AIR TOXICS CONTROLLABILITY STUDY
EPA Contract No. 68-01-7047
Work Assignment No. 26
-------
Although the research described in this report has been
funded wholly or in part by the United States Environmental
Protection Agency, it has not been subject to the Agency's peer
and administrative review process and therefore does not
necessarily reflect the views of the Agency, and no official
endorsement should be inferred.
AIR TOXICS CONTROLLABILITY STUDY
EPA Contract No. 68-01-7047
Work Assignment No. 26
Prepared for:
Eugene Durman - Project Officer
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
P~epared by:
E:H. Pechan & Associates, Inc.
5537 Hempstead Way
Springfield, VA 22151
October 1985
-------
EXECUTIVE SUMMARY
This study was designed to examine existing personal
exposure to 15 different air toxic compounds in five major urban
areas and to estimate how exposure levels are likely to change
between now and 1995 under several different scenarios concerning
future regulatory programs for criteria pollutants and NESHAPs.
Costs associated with each of these regulatory scenarios are
estimated to allow cost effectiveness to be estimated for each
scenario.
Steps taken to perform the analysis were (1) preparing a
baseline. (1980) emission inventory for each study area, (2)
assembling control technology and cost data for the non-zero
toxic emitters, (3) estimating the effects that control programs
and growth will have on emissions between now and 1995 and how
much these control programs might cost, (4) and using 1995
emissions to estimate the resulting risk from air toxics .
exposure. Human exposure to air toxics in the five study areas
was estimated via the U.S. Environmental Protection Agency's
Human Exposure Model (HEM). HEM combines an annual average
dispersion model, population data at the block group/enumeration
district level, and unit risk factors to estimate the expected
increase in cancer incidence resulting from air toxics exposure.
An analysis was also performed of the maximum hazard to
which people might be exposed in each of the five cities. For
this analysis, only major point sources were modeled, and EPA's
Climatological Dispersion Model (CDM) was used to estimate
ambient concentrations of air toxics. .
The starting point in developing the baseline emissions
inventory used in performing these two analyses was the National
Acid Precipitation Assessment Program (NAPAP) inventory for 1980.
The NAPAP inventory was chosen because it is considered to
represent the best available detailed inventory of emissions on a
national scale developed to date. To further improve the NAPAP
inventory, specific air toxics emissions information was obtained
from EPA Emission Factor Documents, Source Assessment Documents,
and local air toxics surveys.
The emission control data base characterized three
important elements: (1) the degree to which existing (1980
baseline) emissions are controlled and their associated costs of
control, (2) the degree to which both existing and new sources
will be controlled in the projection year (1995) and as a result
of. new regulatory requirem~nts imposed between 1980 and 1995, and
(3) the maximum level of control potentially achievable for each
source of toxic emissions. A key assumption used in assembling
this data base was that species of the criteria pollutants are
controlled to the same degree as the criteria pollutant itself.
11
-------
Thus, the values indicating the degree of control of particulate
matter (PM) or volatile organic compound (VaC) emissions were
applied to each toxic species of PM and vac.
The existing degree of control for stationary sources was
determined for each source type using information from the 1980
NAPAP emission inventory. For mobile sources, cost effectiveness
values were calculated based on annual per vehicle costs of the
Federal Motor Vehicle Control Program, Inspection and Maintenance
Programs, and lead phasedown.
Three pieces of information were used to project changes
in future emissions: (1) the rate at which old equipment will be
replaced with new, less polluting equipment, (2) the rate at
which the industry (or emissions source category) is expected to
experience growth in a region, and (3) the constraints that
existing and future environmental .regulations impose on sources
to reduce emissions. Emission control programs considered in
establishing emission constraints included state and local
regulations, New Source Performance Standards (NSPSs), and
NESHAPs.
The regulatory scenarios examined are described in Table
1 along with their impacts on expected excess cancer incidence.
This table gives an overview, and results are discussed in more
detail in Chapters V and VI. Chapter V discusses the exposure.
modeling performed and the results found. Products of incomplete
combustion (PIC) and chromium are the two major contributors to
increased incidence levels under~existing conditions (see Table
V.4 and V.5). In total, PIC emitted by road vehicles and
residential wood stoves and fireplaces represents 81 percent of
all incidence in the five study areas. An additional 10 percent
of total incidence results from chromium emissions. Ethylene
oxide, benzene and formaldehyde are the only other pollutants
which contribute more than 1 percent to the total incidence.
Chapter VI focuses on identifying the maximum risk of increased
cancer incidence to which people might be exposed in each of the
five study areas in search of possible "hot spots." In most
instances it was found that risk hot spots are caused by single
sources, although these sources may emit more than one toxic
compound.
111
-------
Table 1
Sunmary of Estimated Annual Cancer -Incidence
by Scenario*
Estimated Percentage Addi ti onal
Annual Reducti on Percentage
Incidence from Existing Reduction
Existing Conditions (1980) 802
1991) Scenarios
1. Criteria Pollutant Program
Plus Existing NESHAPs 584 27.2% 27.2%
1 a. With New NESHAP Initiatives
A. For chranilll1 only 557 30.6% 4.6%
B. All pollutants 531 33.7% 4.5%
2a. With most stringent contral s
on new sources 529 34.0% 0.6%
2b. With most stringent contral s
on road v ehicl es 528 34.1$ 0
2c. With most stringent controls
on replacement sources 497 38.0% 5.9%
3 . With most stringent control s
on all (new, replacement
and existing) sources 455 43.2% 8.4%
f Incidence nlll1bers presented above aSSlDle constant exposure to pollutant
levels during a 70 year period. Cancer unit risk values applied were
obtained fran EPA' s Carcinogen Assessnent Group.
lV
-------
CONTENTS
TAB LES . . .
FIG URES
. . . . . .
. . . . . . .
. . .
. . .
. . . . .
. . .
. . . .
. . .
. . . . . .
. . .
. . .
. . . .
I INTRODUCTION. . . . . . . . . . . . . . . . . . . . . .
A. BACKG ROU ND AND PU RPOS E . .. . . . . . . . . . . . .
B. KEY MODELING ASSUMPTIONS AND CAVEATS. . . . . . . .
C. REPORT ORGANIZATION. . . . . . . . . . . . . . . . .
II DEVELOPMENT OF BASELINE INVENTORY. . . . . . . . . . .
A. INTRODUCTION. . . . . . . . . . . . . . . . . . . .
B. COMPARISON OF PECHAN AND 35 COUNTY STUDY
INVENTORIES. . . . . . . . . . . . . . . . . . . .
1. Toxic Emissions from POTWs. . . . . . . . . . . .
2. Toxic Emissions from Waste Oil Burning. . . . . .
C. SOU RCE ASSESSMENT DOCUQMENTS. . . . . . . . . . . . .
D. EMISSION FACTORS. . . . . . . . . . . . . . . . . .
E. LOCAL AIR TOXIC SURVEYS. . . . . . . . . . . . . . .
F. EMISSION ESTIMATES PROVIDED BY RADIAN CORPORATION. .
1. Data and Sources. . . . . . . . . . . . . . . . .
- 2. Residential Wood Combustion. . . . . . . . . . .
a. Fireplace Emission Estimate Methodology. . . .
b. Primary Heating Emission Estimate
Methodology. . . . . . . . . . . . . . . . .
c. Secondary Heating Emission Estimate
Methodology. . . . . . . . . . . . . . . . .
G. POLLUTANT-BY-POLLUTANT SUMMARY. . . . . . . . . . .
1. Ar se ni c. . . . . . . . . . . . . . . . . . . . .
2. Benzene. . . . . . . . . . . . . . . . . . . . .
3. Carbon Tetrachloride. . . . . . . . . . . . . .
4. Chloroform. . . . . . . . . . . . . . . . . . .
5. Ch rom i urn . . . . . . . . . . . . . . . . . . . .
6. Ethylene Dichloride. . . . . . . . . . . . . . .
7. Ethylene Oxide. . . . . . . . . . . . . . . . .
8. Formal dehy de . . ~ . . . . . . . . . . . . . . .
9. Gasoline Vapors. . . . . . . . . . . . . . . . .
'0. Lea d . . . . . . . . . . . . . . . . . . . . . .
11. N i ekel . . . . . . . . . . . . . . ~ . . . . . .
12. Perchloroethylene. . . . ... . . . . . . . . . .
13. Products of Incomplete Combustion. . . . . . . .
1~. Trichloroethylene. . . . . . . . . . . . . . . .
15. Vinyl Chloride. . . . . . . . . . . . . . . . .
v
.E.ag~
vii
ix
1
1
3
6
7
7
8
8
8
10
10
11
11
1 1
13
13
1 ~
1 4
15
16
16
16
20
20
20
23
23
23
23
28
28
28
28
34
-------
CONTENTS (continued)
III CONTROL TECHNIQUES EVALUATION. . . . . . . . . . . . .
A. DEVELOPMENT OF EMISSION CONTROL DATA BASE. . . . . .
B. USE OF CONTROL TECHNIQUES DATA BASE. . . . . . . . .
1. General Use. . . . . . . . . . . . . . . . . . .
2. Mobile Sources. . . . . . . . . . . . . . . . . .
IV EMISSIONS AND COST PROJECTIONS. . . . . . . . . . . . .
A. NEW SOURCE GROWTH RATES. . . . . . . . . . . . . . .
B. ESTIMATES OF FUTURE UTILITY EMISSIONS. . . . . . . .
C. ESTIMATED EFFECTS OF ALTERNATIVE CONTROL PROGRAMS. .
1. Development of Analytical Framework. . . . . . .
a. Equipment Replacement Rate File. . . . . . . .
b. New Source Growth Rates. . . . . . . . . . . .
c. Constraint File. . . . . . . . . . . . . . . .
2. Development of Emission Constraints Data. . . . .
3. Effects of Current and Future Regulatory
Requirements on Toxic PM and VOC Emissions
and Control Costs. . . . . . . . . . . . . . .
V EXPOSURE MODELING. . . . . . . . . . . . . . . . . . .
A. MODELING METHODS. . . . . . . . . . . . . . . . . .
. B. RESULTS. . . . . . . . . . . . . . . . . . . . . . .
C. SENSITIVITY ANALYSIS. . . . . . . . . . . . . . . .
D. SUMMARY AND CONCLUSIONS. . . . . . . . . . . . . . .
VI HOT SPOT ANALYSIS. . . . . . . . . . . . . . . . . . .
A. MODELING METHODS. . . . . . . . . . . . . . . . . .
B. RES U L TS . . . . . . . . . . . . . . . ... . . . . . .
1. Ph oe ni x . . . . . . . . . . . . . . . . . . . . .
2. Los Angeles. . . . . . . . . . . . . . . . . . .
3. Baton Rouge. . . . . . . . . . . . . . . . . . .
4. Bal ti mo r e . . . . . . . . . . . . . . . . . . . .
5. Philadelphia~ . . . . . . . . . . . . . . . . . .
c. SU MMAR Y . . . . . . .' . . . . . . . . . . . . . . . .
ABBREVIATIONS AND ACRONYMS. . . .
. . . . . .
. . .
. . . .
REFERENCES. .
. . . . .
. . .
. . .
. . . .
........
vi
.E,gg~
36
36
40
40
40
43
43
43
43
45
45
47
48
49
58
76
76
80
91
95
96
96
96
96
98
102
102
102
109
110
112
-------
.N.YID.Qg.r
I.1
I.2
II. 1
II.2
II. 3
II.4
II.5
II.6
II. 7
II.8
II.9
II.10
II. 11
II. 12
II. 1 3
II. 1 4
II. 1 5
II. 16
III. 1
III.2
III. 3
IV .1
IV.2
IV.3
IV.4
IV.5
IV.6
IV.7
Iv.8
V.1
V.2
°V.3
V.4
TAB LES
Study Areas. . . . . . . . . . . . . . . . . .
. . .
Compounds Included in the Study. . . . . . . . . . .
Emission Estimates Provided by Radian Corporation. .
Arsenic Baseline Emissions. . . . . . . . . . . . .
Benzene Baseline Emissions. . . . . . . . . . . . .
Carbon Tetrachloride Baseline Emissions. . . . . . .
Chloroform Baseline Emissions. . . . . . . . . . . .
Chromium Baseline Emissions. . . . . . . . . . . . .
Ethylene Dichloride Baseline Emissions. . . . . . .
Ethylene Oxide Baseline Emissions. . . . . . . . . .
Formaldehyde Basel ine Emissions. . . . . . . . . . .
Gasoline Vapors Baseline Emissions. . . . . . . . .
Lead Baseline Emissions. . . . . . . . . . . . . . .
Nickel Baseline Emissions. . . . . . . . . . . . . .
Perchloroethy lene Basel ine Emissions. . . . . . . .
Products of Incomplete Combustion, Benzo(a)pyrene
Baseline Emissions. . . . . . . . . . . . . . . .
Trichloroethylene Baseline Emissions. . . . . . . .
Vinyl Chloride Baseline Emissions. . . . . . . . . .
Generic Cost Effectiveness Values for New Source
VDe Emissions. . . . . . . . . . . . . . . . . . .
Generic Cost Effectiveness Values for New Source
PM Emissions. . . . . . . . . . . . . . . . . . .
Cost Effectiveness and Emission Reduction Values
for Control of Mobile Source PM and VOC
Emissions. . . . . . . . . . . . . . . . .
. . . .
Population and Economic Growth Rates: State Two-
digit SIC Allocated to Selected SMSAs. . . . . . .
Emissi on Proj ecti on Modul e Records. . . . . . . . .
Emission Constraints. . . . . . . . . . . . . . . .
Description of Regulatory Scenarios. . . . . . . . .
Summary of PM Emissions Reduction and Associated
Costs of Control by Scenario. . . . . . . . . . .
Summary of VOC Emissions Reductions and Associated
Costs of Control by Scenario. . . . . . . . . . .
PM and VOC Emission Control Cost Increases in all
F i v e Ar e as. . . . . . . . . . . . . . . . . . . .
Major Source Categories Studied. . . '0 . . . . . . .
Unit Risk Values Used in the Air Toxics
Controllability Study. .0. . . . . . . . . . . . .
Minimum Emission Limits Used for Identifying Major
Point Sources. . . . . . . . . . . . . . . . . . .
Summary of Estimated Annual Cancer Incidence by
See na r i 0 . . . . . . . . . . . . . . . . . . . . .
Annual Incidence Under Existing Conditions: Top
Five Source Category-pollutant Combinations. . . .
vii
E.sgg
2
5
12
17
18
19
21
22
24
25
26
27
29
30
31
32
33
35
38
39
42
44
50
51
57
71
7°2
73
75
77
79
81
86
-------
R~mQ~I
V.5
V.6
V.7
V.8
V.9
TABLES (continued)
Incidence Under Existing Conditions: Top Five
Source Category-pollutant Combinations. . .
Summary of Estimated Annual Cancer Incidence by
Scenario. . . . . . . . . . . . . . . . . .
. . .
. . .
Expected Annual Cancer Incidence by Pollutant
Assuming Increasing Regulation. . . . . . . . . .
Expected Annual Cancer Incidence by Source
Category Assuming Increasing Regulation~ . . . . .
Sensitivity of Estimated Annual Cancer Incidence
to Assumptions About the Percentage of Chromium
Emissions that is Hexaialent . . . . . . . . . . .
viii
E~g~
87
89
90
92
94
-------
.N.Y1!.1.Q~.L
IV.1
IV.2
IV.3
IV.4
IV.5
IV.6
IV.7
IV.8
IV.9
IV . 1 0
IV . 1 1
IV . 1 2
IV . 1 3
IV . 1 4
V. 1
V.2
V.3
VI. 1
VI.2
VI. 3
VI.4
VI.5
VI. 6
VI. 7
VI. 8
VI. 9
VI. 10
FIG URES
Schematic of Regulatory Impact Model. . . . . . . .
Constraint Code Sheet for Input to RIM: . . . . . .
Total Uncontrolled afid Controlled Toxic ~M
Emissions by Scenario - Phoenix. . . . . . . . . .
Total Uncontrolled and Controlled Toxic PM
Emissions by Scenario - Los Angeles. . . . . . . .
Total Uncontrolled and Controlled Toxic PM
Emissions by Scenario - Baton Rouge. . . . . . . .
Total Uncontrolled and Controlled Toxic PM
Emissions by Scenario - Baltimore. . . . . . . . .
Total Uncontrolled and Controlled Toxic PM
Emissions by Scenario - Philadelphia. . . . . . .
Total Uncontrolled and Controlled Toxic PM
Emissions by Scenario - All Areas. . . . . . . . .
Total Uncontrolled and Controlled Toxic vac
Emissions by Scenario - Phoenix. . . . . . . . . .
Total Uncontrolled and Controlled Toxic vac
Emissions by Scenario - Los Angeles. . . . . . . .
Total Uncontrolled and Controlled Toxic vac
Emissions by Scenario - Baton Rouge. . . . . . . .
~otal Uncontrolled and Controlled Toxic vac
Emissions by Scenario - Baltimore. . . . . . . . .
Total Uncontrolled and Controlled Toxic vac .
Emissions by Scenario - Philadelphia. . . . . . .
Total Uncontrolled and Controlled Toxic vac
Emissions by Scenario - All Areas. . . . . . . . .
Total Annual Incidence by Scenario. . . . . . . . .
City-specific Analysis - Existing Conditions. . . .
City-specific Analysis - Existing Conditions. . . .
Key Air Toxics Point Source Hazards - Phoenix. . . .
Hazards from Major Point Sources - Phoenix. . . . .
Key Air Toxics Point Source Hazards - Los Angeles. .
Hazards from Major Point Sources - Los Angeles. . .
Key Air Toxics Point Source Hazards - Baton Rouge. .
Hazards from Major Point Sources - Baton Rouge. . .
Key Air Toxics Point Source Hazards - Baltimore. . .
Hazards from Major Point Sources - Baltimore. . . .
Key Air Toxics Point Source Hazards - Philadelphia.
Hazards from Major Point Sources - Philadelphia. . .
ix
£il.s,e
46
53
59
60
61
62
63
64
65
66
..67
68
69
70
83
84
85
97
99
100
101
103
104
105
106
107
108
-------
I INTRODUCTION
A. BACKGROUND AND PURPOSE
This study was designed to examine existing personal
exposure to 15 different air toxic compounds in five major urban
areas and"to estimate how exposure levels are likely to change
between now and 1995 under several different scenarios concerning
future regulatory programs for criteria pollutants and NESHAPs.
Costs associated with each of these regulatory scenarios are
estimated to allow cost effectiveness to be estimated for each
scenario. The analysis concentrated on five cities so that air
toxic emission estimates would be as accurate as possible. These
cities were chosen to represent a cross section of potential
problem types: predominant mobile sources, mix of industry and
mobile sources, and old industry. The five study areas were
Phoenix, Los Angeles, Baton Rouge, Baltimore, and Philadelphia.
The exact geographic areas within these metropolitan areas
included in the study are listed in Table 1.1.
Work on this study was performed jointly by E.H. Pechan &
Associates, Inc. (Pechan) and Radian Corporation. Pechan was
responsible for preparing the baseline emission inventory for the
five study areas. This .inventory was provided to Radian along
with industry specific growth rates for the five study areas.
R~dianthen assembled control technology and cost data for the
source categories which were non-zero toxic emitters. Radian
also estimated the effects that various control programs and
growth would have on emissions between now and 1995 and how much
these control programs might cost. Radian analysts used this
information to produce estimates of the percentage change in
particulate and organic emissions between the base year (1980)
and 1995. This information was transferred to Pechan and was
used to estimate 1995 emissions and the resulting risk from air
toxics exposure. Human exposure to air toxics in the five study
areas was estimated via the U.S. Environmental Protection
Agency's (EPA) Human Exposure Model (HEM). The HEM combines an
annual-average dispersion model, population data at the block
group/enumeration district level, and unit risk factors to
estimate the expected increase in cancer incidence resulting from
air toxicsexposure.
An analysis was also performed of the maximum risk to
which people might be exposed in each of the five cities. For
this analysis, only major point sources were modeled, and EPA's
Climatological Dispersion Model (CDM) was used to estimate
ambient concentrations of air toxics.
Fifteen different compounds were studied in this
analysis. These compounds were arsenic, benzene, carbon
tetrachloride, chloroform, chromium, ethylene dichloride,
1
-------
Table 1. 1
Study Areas
1980
MetroDolitan Areas Population Counties
Phoenix, Arizona 1,509,000 Maricopa
Los Angel es, California 10,867,000 Los Angeles
Orange
Riverside*
San Bernardino *
Baton Rouge, Louisiana 383,000 East Baton Rouge Parish
West Baton Rouge Parish
Bal timore, Maryland 1,438,000 Bal timore Ci ty and County
Phil adel phi a , Pennsylvania 1 ,688,060 Phil adelphia
*
Portions in the South Coast Air Basin
2
-------
ethylene oxide, formaldehyde, gasoline vapors, lead, nickel,
perchloroethylene, products of incomplete combustion,
trichloroethylene, and vinyl chloride. Although there are many
more compounds present in urban atmospheres than those listed
above, the ones studied are believed to be the most important in
terms of cancer risk.
B. KEY MODELING ASSUMPTIONS AND CAVEATS
In any study, the results are somewhat reflective of the
modeling techniques used and the assumptions made about how
changes in important variables will affect the end result. This
study is no different; therefore, it is useful at the outset for
the reader to understand how the study was performed so that the
results can be viewed in the proper context. .
Some of the more important assumptions and caveats
associated with this study are as follows:
1. Personal exposure to air toxics was estimated using
annual-average concentration estimates and it was assumed
that exposures occur where people reside. In addition,
only outdoor exposures were modeled. Thus, this
methodology ignors the fact that people move throughout
the urban area, travel outside the urban area, and are
indoors a majority of the day. Because exposures were
simulated over a 70-year period, it is unclear how much
this restricted modeling methodology affects the study
results.
2. The study relied sOlely on quantitative estimates of
cancer risk associated with inhalation of ambient air.
Acute and sub-chronic effects were not included, and
cancer cases associated with exposure routes other than
. inhalation of ambient air were not quantified.
3. Only 15 compounds were included in this study, although
monitoring studies have shown that urban atmospheres
typically contain many more pollutants.
4. Annual-average emission estimates were used to estimate
concentrations of air toxics. Thus, the study focused on
routine, continuous emissions. Accidental releases were
not modeled.
5. Unit risk factors employed ln this study represent the
chance of contracting cancer from a lifetime (70 years)
exposure to a given concentration of that pollutant. The
carcinogenic potency estimates used in this study were
developed by EPA's Carcinogen Assessment Group (CAG).
6. Cancer incidence estimates are presented in this report
3
-------
for existing conditions (1980) and for a number of 1995
scenarios. These incidence estimates are based on the
assumption that emission levels for each scenario remain
constant for a 70-year period. In reality, emissions
will vary from year-to-year. Pechan has not attempted to
capture these variations in this analysis.
7. Study results show that products of incomplete combustion
(PIC) and chromium are the major contributors to cancer
incidence in the five study areas. There is considerable
uncertainty in the methods used to estimate cancer
incidence for each of these two compounds.
For PIC, benzo(a)pyrene (B(a)P) emission data were used
as a surrogate to estimate cancer incidence associated
with this compound. There are many limitations to this
approach. For example, the proportion of carcinogenic
activity attributable to B(a)P in PIC mixtures is known
to vary among source categories and sometimes within a
source category. Thus approximation was deemed useful
and necessary for the study of PIC, however.
For chromium, the major uncertainty results from the
difficulty in determining whether emissions are in the
trivalent or hexavalent form. Only the hexavalent form
has been proven carcinogenic; there is insufficient
evidence to determine if the trivalent form is also
carcinogenic. Since there are limited data on the ratio
of trivalent to hexavalent chromiu~ in emissions, it was
necessary to make an assumption. This analysis assumes
that total chromium is as carcinogenic as the hexavalent
form.
8. In our analysis of the effects of control techniques on
toxic emissions, it has been assumed that toxic emissions
are controlled to the same extent as criteria pollutants.
Thus, if a control technique is expected to reduce
particulate emissions by 90 percent, it is assumed here
that particulate toxics are also reduced by 90 percent.
In order to employ this assumption in the analysis, it
was necessary to group the 15 compounds under study into
organics and particulates (or metals). This
classification is shown in Table 1.2. It should be noted
that if emissions of some compounds are reduced more than
others by a control techniq~e, the difference will not be
captured in the analysis.
9. In assessing cancer risk within an urban area, each of
the compounds under st~dy has been analyzed individually.
Any possible synergistic health effects of these
compounds have been ignored. .
4
-------
Ta bl e 1. 2
Compounds Included in the Study
.Q.rg,g.D.1.Q~
.f.a.r.t.1.Q'y.J..B.t~~
Benz ene
Arsenic
Carbon Tetrachloride
Chromi urn
Chloroform
Lead
Ethylene Dichloride
Nickel
Ethylene Oxide
Products of Incomplete Combustion
Formal dehy de
Gasol ine Vapors
Perchloroethylene
Trichloroethylene
V iny I ChI ori de
5
-------
10. Sources included in the exposure modeling data set for
each study area were limited to those in the counties
under study. Therefore, while contributions from these
sources to areas outside the county boundaries were
considered, contributions of sources located outside the
county boundaries to air toxic concentrations within the
study areas were not. Of the five study areas, it is our
sense that the Baton Rouge results are most affected by
this assumption. There are some significant toxic
emitters in Plaquemine and Geismar, Louisiana, which were
not included in this analysis, even though they may have
some effect on air toxic concentrations in the study
area. .
11. Modeling results presented in this report for exposure to
gasoline vapors do not include self-service exposures at
service stations. Thus, the modeling results for this,
pollutant will not be directly comparable with those
presented in the gas marketing study (U.S. EPA, 1984k).
C. REPORT ORGANIZATION
Besides this introduction, there are five chapters in
this.analysis. Chapter II presents t~e methods used to develop
the baseline (1980) emission inventory and summarizes the
pollutant specific emission estimates for each study area. .
Chapter III discusses the available control techniques for each
source type and identifies the cost and effectiveness of each
control option. Emissions and cost projections to 1995 for six
different control scenarios are presented in Chapter IV. The
different control scenarios discussed include the criteria
pollutant regulatory program, NESHAPs, and other more stringent
controls.
. Exposure modeling methodologies and results are presented
in Chapter V. The presentation of results includes the total
incidence across all study areas, the expected reductions in
incidence via various contrOl. programs, and the source categories.
and pollutants driving the reductions. City specific analyses
are also ~resented in Chapter V.
Chapter VI presents the results of an analysis to
estimate the maximum risk from major point source contributions
in each study area. This chapter snows the relationship between
hazard level and population, shows the spatial variation in
hazard within each study area, and identifies whether points of
maximum hazard are a single or multiple source problem.
Detailed data summaries are presented in the appendices.
6
-------
II DEVELOPMENT OF BASELINE INVENTORY
A. INTRODUCTION
The starting point in developing the baseline inventory
was the National Acid Precipitation Assessment Program (NAPAP)
inventory for 1980. The NAPAP inventory was chosen because it is
considered to represent the best available detailed inventory of
emissions on a national scale developed to date. The NAPAP
inventory is similar to EPA's National Emissions Data System
(NEDS) but it has been improved substantially by incorporation of
the latest available emission factors, substituting data from the
Northeast Corridor Regional Modeling Project (NECRMP) and recent
state submissions, and incorporating electric utility data
compiled by Pechan.
To further improve the NAPAP inventory, specific air
toxic emissions information was obtained from EPA Emission Factor
Documents, Source Assessment Documents, and local air toxic
surveys. If more than one data source was available for a
pollutant and source category the following criteria, listed in
order of priority, were applied to estimate air toxic emissions:
1. Plant specific emission information from local air toxic
surveys or source assessment documents
2. Emission factors
3. National or statewide emission estimates apportioned to
counties
The inventories developed using these procedures were
reviewed by the appropriate local agency for each study area.
Corrections were then made to the air toxic inventories based on
comments received from these local agencies. The plant specific
information needed was facility name, control equipment, control
effectiveness, location (Universal Transverse Mercator -- UTM --
coordinates), pollutant, emissions, process (by Source
Classification Code -- SCC), and stack parameters (height,
diameter, temperature, and flow).
All of these data items were necessary in order to meet
the multiple needs of emission inventory users." Locations,
emissions, and stack data were need~d for dispersion modeling.
Controlled and uncontrolled emissions and current control
equipment data were needed to estimate the effect of future
regulations and to estimate future control costs.
7
-------
B. COMPARISON OF PECHAN AND 35 COUNTY STUDY INVENTORIES
The methodology used to estimate baseline (1980) toxic
emissions paralleled that used in the 35 County Study (Versar,
1984), with some notable exceptions. Pechan started with the
1980 NAPAP inventory, which includes the NECRMP point source data
for Philadelphia and Baltimore. This differs from the 35 County
Study, which used the 1981 NEDS snapshot file as its starting
point. For some areas, the NAPAP inventory and the 1981 NEDS
snapshot file may be identical, except that Pechan utility data
have been substituted in NAPAP for that reported to NEDS by the
sta tes.
The five study areas were selected on the basis of data
availability and the ability to reflect a range of urban
characteristics (geographic distribution, industrialization, and
population). The number of toxic pollutants to be included in
the emission inventory was reduced from the 21 used in the 35
County Study to 14 and lead was added. These 15 pollutants were
chosen to reflect the highest risk pollutants. Choosing a small
number of study areas and only examining the highest risk
pollutants allowed Pechan to obtain more local information and to
consult additional sources in detail for each study area. .
Pechan adopted the 35 County study methodologies for
estimating emissions from Publ icly Owned Treatment Works (POTWs)
and waste oil burning. Brief descriptions of those methodologies
are presented below. . .
1. Toxic Emissions from POTWs
The 35 County study methodology for calculating emissions
from POTWs uses a multistep process. First, EPA's Industrial
Facilities Discharge (IFD) file was used to identify POTWs known
to handle industrial discharges. The following plant specific
data were extracted from the IFD file for each POTW: county,
state, total flow, level of treatment (primary, secOndary/: or
tertiary), percen~age industrial contribution, and type 0
industry discharging into the POTW (by SIC code). The POTWs were
next sorted into 13 categories based on industrial dischargers
and level of treatment. The 35 County study then derived seven
prototype plants and two equations to be used to calculate annual
POTW air toxic emissions. . Industrial discharges and level of
treatment were used to determine the appropriate equation and
prototype plant to be applied.
2. Toxic Emissions from Waste Oil Burning
The following is the methodology used by the 35 County
study to calculate toxic emissions from waste oil burning.
First, each state's total residual oil consumption (TRESD) for a
state was obtained from NEDS. If TRESD was not zero, the
8
-------
following two equations were used to calculate waste oil burned
based on residual oil consumption (residential (RES-RESD),
commercial/institutional (CI-RESD), and industrial (IND-RESD».
Residential, institutional, and commercial (RIC) boilers:
RIC = (RES-RESD + CI-RESD + IND-RESD) in countv (gals) x
(gals) TRESD for state (gals)
-2l waste oil burned
100 x in state (gals)
(1)
Industrial boilers (IND):
IND = (RES-RESD + CI-RESD + IND-RESD) in countv (gals) x
(gals) TRESD for state (gals)
-I2 x waste oil burned
100 in state (gals)
(2)
If TRESD was zero, waste oil burned was estimated by
replacing residual oil consumption with distillate oil
consumption (RES-DIST, CI-DIST, and IND~DIST). Using RIC and IND
from equations (1) and (2), metal emissions and organi'c emissions
were then calculated using equations (3) and (4) below. .
Me tal Emi ssi ons (MEMIS):
MEMIS -(RIC + IND) gals x emission rate (gm/m3)
(kkg) -
x 3391.36 (10-6) kkg oil
gal waste oil
x -12 x metal concentration (ppm) x 2.7 (10-6) kkg
100 kkg oil
( 3)
Organic Emissions (OEMIS):
o E MI S = [( R I C x 0.0 1) + (I N D x 0.001)] gal s x 3391.36 (1 0 - 6 )
(kkg)
kkg oil
gal waste oil
x organic concentration (ppm) x 160 (10-6)
kkg
kkg oil
The numbers associated with kkg oil/gal waste oil and
kkg/kkg oil are unit conversion factors.
9
-------
C. SOURCE ASSESSMENT DOCUMENTS
Source assessment documents have been developed by EPA to
provide information on significant emitters of an air pollutant
to help determine whether regulation is needed and how effective
controls might be. In these documents each source category is
assessed separately and emission information is provided either
in the form of plant specific emissions or nationwide emission
rates. Plant specific information usually includes plant name,
location (longitude-latitude or UTM coordinates), stack
parameters, pollutant, and emissions. . Plant specific data
obtained from these documents were used in preference to those in
the NAPAP file.
Nationwide emission rates were used for source categories
for which plant specific or area specific emissions could not be
found. These source categories were usually area sources (grain
fumigation, unidentified, or miscellaneous sources). The
nationwide average emission estimates were apportioned to
counties by population. If regional or state emission estimates
were provided, these data were used instead of apportioning the
nationwide average estimate. .
D. EMISSION FACTORS
Emission factors for air toxic compounds were developed
from a number of different sources including EPA emission factor
documents, voe species profiles, and air pollution periodicals.
EPA emission factor documents are for individual pollutants and
contain uncontrolled emission factor estimates for numerous
sources. However, emission factor documents were only available
for carbon t~trachloride, chloroform, chromium, ethylene
dichloride, formaldehyde, and nickel. .
The PEDCo paper "Trace Emissions from Coal and Oil
Combustion" provided trace metal (arsenic, chromium, lead, and
nickel) emission estimates for various coal and oil fired boilers
(PEDCo, 1982). Emission factors were provided for uncontrolled
sources and for specific types of control equipment at each
source. Formaldehyde emission factors from combustion sources
were obtained from the EPA formaldehyde emission factor document
(U.S. EPA, 1984e). These emission factors also reflect
uncontrolled and controlled estimates.
Mobile sources emit benzene. (Gray, 1982), B(a)P (PAH,
1983), formaldehyde (Carey, 1981) and lead (U.S. EPA, 1984g).
Emission factors for lead were calculated by multiplying the
fraction of the vehicle fleet that uses leaded gas by its travel
fraction for each vehicle year, then summing each product. The
same calculation was performed for the fraction of the vehicle
fleet that uses unleaded gas. Each sum was multiplied by its
10
-------
respective lead content. (Gasoline lead contents used in this
analysis were 1.38 grams per gallon for leaded fuel and 0.014
grams per gallon for unleaded fuel.) These two concentrations
were added together and divided by the fleet miles per gallon to
obtain an emission factor in grams per mile. Emission factors
for benzene, B(a)P, and formaldehyde were obtained directly from
the referenced sources. B(a)P emission factors for non-mobile
sources are the same as those used in the 35 County Study.
Uncontrolled lead emission factors for stationary sources
were obtained from AP-42 (U.S. EPA, 1984a). Source categories in
AP-42 were matched with SCCs to develop an emission factor file
for lead emissions. Where controlled emission factors were
known, the appropriate control equipment code was also assigned.
E. LOCAL AIR TOXIC SURVEYS
Local agency air toxic emission estimates were used to
estimate air toxic emissions for a number of plants in Los
Angeles, Baltimore, and Philadelphia. Los Angeles data were
available from a 1982 survey of approximately 1,800 facilites.
Results of this survey are available in the South Coast Air Basin
. (SCAB) report (SCAQMD, 1983). Data for two significant air toxic
emitters in Baltimore County were received from the Maryland Air
Management Administration. Philadelphia air toxic emission
estimates were taken from the Philadelphia Air Management.
Office's computer data base - PIPQUIC. The PIPQUIC data file
provided information on plant name, location (UTM coordinates1,
stack parameters, pollutants, and emissions. The SCAB report and
the Maryland Air Management Administration provided the plant
name, location (street and city), pollutants, and emissions.
These data sets were considered more accurate than the data
present in NAPAP. Therefore, the three data sets were compared
to NAPAP and the applicable changes were made. Each plant's
emissions were assigned to the appropriate SCC code and used to
override emission factors.
When the PIPQUIC data were compared to NAPAP, 33 PIPQUIC
plants were not found in NAPAP. For these plants, new records
were created and entered into the NAPAP file.
F. EMISSION ESTIMATES PROVIDED BY RADIAN CORPORATION
1. Data and Sources
Radian Corporation provided Pechan with baseline emission
estimates for dry cleaners (VOC), refractory manufacturing (TSP),
residential wood combustion (TSP and VOC), and sterilizers
(ethylene oxide). These emission estimates are listed in Table
11.1. The dry cleaning vac estimates were obtained directly from
11
-------
Categorv
PCE Dry Cleaning
Sterilizers (Ethylene Oxide)
Residential Wood Combustion:
Fireplaces
Woodstoves
I-'
I\)
Categorv
Refractory Manufacturing
Residential Wood Combustion:
Fireplaces
Wood Stoves
* not available
Table, II. 1
Emissi~n Estimates Provided by Radian Corporation
VOC Emissions CTPY)
Phoenix SCAB Baton Rouge Baltimore Philadephila
911 ° * 443 338.4
7.8 44 2.8 9.7 11 .4
923 6,968 461 820 243
6 , 177 32,832 1,639 5,180 2,457
TSP Emissions CTPY)
Bal timore
Phoenix SCAB Baton Rouge uncontrolled controlled Philadelphia
* * * 5,192.1 236.06 *
1 ,008 7,295 475 866 260
2,592 13 ,77 1 685 2,234 1 ,040
-------
state agencies. The data for refractory manufacturing were
extracted from source assessment documents and. represent two
model plants located in Baltimore. Sterilizer estimates include
only ethylene oxide emissions from hospital usage. Actual values
were obtained for Philadelphia and a coefficient was developed;
other areas are based on this value with adjustments made for the
number of hospitals and admissions.
. 2. Residential Wood Combustion
Residential wood combustion emissions were prepared by
OMNI Corporation for Radian Corporation. OMNI's estimates of .
residential wood combustion emissions for each of the five study
areas were divided into three categories: (1) fireplaces used
primarily for ornamental or aesthetic wood burning, (2) primary
heating, done primarily with wood stoves, and (3) secondary
heating, done primarily with modified fireplaces (inserts) and
wood stoves.
a. Fireplace Emission Estimate Methodology
Using the 1980 census data (U.S. DOC, 1982a) in
conjunction with U.S. Bureau of Labor Statistics (.U.S. DOL, 1958)
and U.S. Commerce Department Reports (U.S. DOC, 1975, 1980,
1982b), OMNI estimated at the county level the number of single
family households having one or mor~ fireplaces. OMNI used a
conservative approach, however, in that single family houses
estimated to have two or more fireplaces were assumed .to have
only one operating fireplace. No attempt was made to estimate
the amount of wood burned in fireplaces in multi-family dwelling
units or mobile homes because residential wood fuel use studies
indicate that more than 96 percent of all wood used for
residential wood combustion purposes occurs in single family
dwelling units (Skog and Watterson, 1983).
OMNI applied a utilization factor to each of the
calculated number of installed fireplaces to estimate the number
of operating fireplaces. OMNI estimated only those fireplaces
installed as part of the original construction of single family
homes even though national fireplace sales data (HID, 1984)
indicate that 36 percent to 43 percent of recent fireplace
installations are associated with residential repair and
remodelling projects.
OMNI estimated the number of operating fireplaces used
for ornamental or aesthetic purposes by applying a conversion
factor that assumes a certain percentage of ~xisting operating
fireplaces were modified (i.e., inserts installed) to be used for
secondary heating rather than for aesthetic wood burning
purposes. Fireplace conversion factors were estimated from
residential fuel wood use studies (Skog and Watterson, 1983; EIA,
1982b, 1982c) and a review of regional marketing data of
13
-------
residential wood burning devices.
The average amount of wood burned per dwelling unit
having operating fireplaces was based on a number of studies,
discussions with state and regional foresters, state energy
officials, and commercial/retail wood dealers. OMNI calculated
the total amount of wood (in cords) burned in fireplaces used
primarily for ornamental or aesthetic purposes in a county.
AP-42 fireplace emission factors (U.S. EPA, 1984a) were applied
to the estimated amount of wood burned to estimate TSP and VOC
emissions.
b. Primarv Heating Emission Estimate Methodologv
To estimate primary wood heating emissions, OMNI used
1980 census data to determine the number of households using wood
as a primary hea ti ng fuel. The households answering "wood" are
summarized at the county level in the 1980 Census Survey. Other
studies (Skog and Watterson, 1983; EIA, 1982b, 1982c) have
indicated that the 1980 Census Survey may have underestimated the
number of households using wood as a primary heating fuel by as
much as a factor of 2.5 on a nationwide basis. These studies do
not provide data at a county level, however, and it was OMNI's
opinion that in lieu of more detailed site-specific data the 1980
Census Survey Data should be used and would represent a
conservative estimate of households using wood as a primary
heating fuel.
OMNI made an estimate of the average amount of wood
consumed per primary wood burning household using a variety of
residential fuel woo~ use studies (Skog and Watterson, 19B3; EIA,
1982a, 1982b, 1982c; Marshall, 1981) as well as information. .
derived from discussions with other experts in the wood energy
field. Since there is very little data regarding the average
amount of wood consumed per primary wood burning household for
the areas evaluated, OMNI estimates are to be considered
conservative since the values selected were generally on the low
end of data ranges available from these studies.
For each study area, total amount of wood consumed (in
cords) was estimated for all households using wood as a primary
heating source. It was assumed that all households using wood as
a primary heating fuel used space heating devices that had the
emission performance characteristics of an "average" wood stove.
Therefore, AP-42 wood stove emission factors were applied to this
residential wood combustion class. .
c. Secondarv Heating Emission Estimate Methodologv
As noted above, it was assumed that most of the secondary
or supplemental heating was done primarily with modified
fireplaces or wood stoves. It was recognized that ordinary
14
-------
fireplaces are occasionally used to provide supplemental heat
(primarily for emergencies), but it was OMNI's view that the
emission contribution of this source was insignificant.
The average amount of wood burned per modified fireplace
was based on national and regional residential firewood use
studies (Skog and Watterson, 1983; EIA, 1982b, 1982c; DER, 1982;
NRBP, 1984; USDA, 1980) evaluations of heating demand
characteristics of single family homes by geographic area and
discussions with regional foresters and commercial firewood
dealers.
OMNI assumed, based on its emission testing experience,
that fireplace ihserts essentially have the same emission
characteristics as the average population of both airtight and
non-airtight wood stoves. Therefore, AP-42 wood stove emission
factors were applied to the amount of wood burned in modified
fireplaces for secondary heating purposes to estimate TSP and VOC
emissions.
The number of wood stoves used for secondary heating
purposes was estimated from data in the 1980 Annual Housing
Survey (U.S. DOC, 1983a) and data regarding sales of wood stoves
in the United States (EIA, 1982b). Sales data indicate that the
secondary wood heating device market did not evolve until after
the 1973 Arab oil embargo. Therefore, OMNI estimates reflect
wood stove installations occurring between 1975 and 1980.
The average amount of wood used per secondary heating'
wood stove was determined using the same data sources and
methodology as used for modified fireplaces. AP-42 wood stove
emission factors were us~d to estimate TSP and VOC emissions.
Note that there is considerable uncertainty in the
residential wood combustion emission estimates. Comments
received from the South Coast Air Quality Management District
indicated that the OMNI emission estimates were much higher than
its own estimates. Therefore, OMNI' s .e stimates of the amount of
wood burned per wood burning device were adjusted downward to try
to better account for the local conditions in the South Coast.
G. POLLUTANT-BY-POLLUTANT SUMMARY
This section summarizes the source categories and
emissions of the 14 toxic pollutants and lead. Each pollutant is
di~cussed individually. For each pollutant, baseline emissions
are shown by source category and study area. Brief explanations
of how emissions were estimated are provided for the most
significant source categories.
15
-------
1. Arsenic
Arsenic is emitted by a limited number of sources. Major
sources in the five study areas are boilers, glass manufacturing
plants, and industrial processes. Glass manufacturing and
secondary lead smelting emissions were obtained from plant
specific data. Emission factors were used to estimate emissions
from coal and oil combustion sources (PEDCo, 1982). Emissions
from waste oil burning were calculated using the methodology
presented in the appendix of the 35 County Study (Versar, 1984).
Baseline arsenic emissions are listed by source category for each
study area in Table II.2.
2. Benzene
Benzene is emitted by numerous source types. These
emissions either arise from direct use of benzene or are formed
as by-products during the manufacture of another product. Mobile
sources emit 80 percent of the total benzene emissions in the
five study areas. Emissions from coke oven by-product recovery
are significant, but they are only present in Baltimore.
Emission factors were used to estimate benzene emissions
from mobil e source s (Gray, 1982). POTW emissi ons were cal cuI a ted
using the methodology presented in the appendix of the 35 County
Study (Versar, 1984). Emissions from linear alkylbenzene plants
and coke oven by-product plants are calculated from plant
specific information. Gasoline marketin& emissions were
calculated by apportioning statewide emissions by consumption.
Emissions for all other source categories were calculated by
apportioning nationwide emission estimates by population.
Baseline benzene emissions are listed by source category for each
study area in Table II.3.
3. Carbon Tetrachloride
Carbon tetrachloride is emitted from either the
production of carbon tetrachloride or as a by-product from the
production of other chemicals. Unidentified uses account for 37
percent of the carbon tetrachloride emissions in the five study
areas. To estimate the emissions from this source the nationwide
emissions estimate was apportioned by population. Ethylene
dichloride (EDC) production, which only occurs in Baton Rouge,
accounts for le~ss thCU)~ 1", R~Lgent o~ .the carbon. tetrachloride
jff , emissions. bi.ocal '({ata were use~ to. estimate emissions from EDC
jbr produ~on, pharmaceutical manufacturing, fluorocarbon
f./t'"jt'pr.oduction, and chemical manufacturing, whereas POTW emissions)
I were estimated using the methodology presented in the appendix of
the 35 County Study (Versar,. 1984). Baseline carbon
tetrachloride emissions are listed by source category for each
study area in Table II.4.
/ ,/
£ ~.
:f j-I. 0:/,7
16
-------
Table II.2
Arsenic
Baseline Emissions (TPY)
Los Baton
Source CategorY Phoenix Angeles Rouge Baltimore Philadelphia
Utility Boilers 0.2 1.3 0.4 0.8
Industrial Boilers 0.7 0.2 0.8 1.2
Secondary Lead Smelters 0.0 0.3
Residential Wood Combustion 0.0
Glass Manufacturing 2.4
Waste Oil Burning 0.1 0.0 0.1 0.1 0.2
Total 0.3 2.0 0.6 3.7 2.2
17
-------
Table II. 3
Benzene
Baseline Emissions (TPY)
Los Baton
Source Categorv Phoenix Angeles Rouge Baltimore Philadelphia
Utility Boilers 0.4 5.8 0.5
Industrial Boilers 0.4 18.1
Coke Ovens -
By-Product Recovery 0.0 1570
Petroleum Refinery
Fugitives 638.5 85.9 155
Gasoline Marketing 70.6 25.0 8.5 23.6 19.6
Benzene Fugitives 240.0 178.6 60 93.6
Chemical Manufacturing - Misc 117.4
Linear Alkylbenzene Vents 280
Ethylbenzene/Styrene 0.8
Motor Vehicles - Off-Hwy 62.3 372.2 13.9 38.1 52.2
Motor Vehicles - Hwy 1 ,468 10,577 360 1 , 160 861
Gasoline-Powered Vessels 0 63.1 2.7 13.2 4.2
IC Engines/Turbines 2.1
P01Ws 1.0 7.8 5.6 15.4
Residential - Commercial
Gas Furnaces 1.8 31.2 1.2 3.4 6.8
Incineration 7.1 2.3 1.9 0
Agricultural Burning 6.0
Sol vent Use - Benzene 1.1 6.9 0.3 1 .1 1.3
Benzene Storage 12.2 10.5 1.8 10.5
Benzene Handling 7.5 46.7 1.9 7.2 8.4
Total 1 ,620 12,154.6 664.8 3 , 166 1 ,246
18
-------
Table I1.4
Carbon Tetrachloride
Basel ine Emissions (TPY)
Los Baton
Source Categorv Phoenix Angeles Rouge Baltimore Philadelchia
EDC Production* 2.38
Pharmaceutical
Manufacturing 17.4
Fluorocarbon Production 3.1
Chanical Manufacturing 4.0 0.3
P01W's 4.6 10.5 -
Grain Fumigation 52.8 0.0 164.0 73.7 107.3
Unidentified Uses 185.0 0.0 46.9 176.0 206.0
Total 237.8 11.7 213.6 260.2 330.7
* This is the emission estimate for Formosa Plastics. Emissions were
calculated by the Louisiana Depar.tment of Environmental Quality.
19
-------
4. Chloroform
Chloroform is mainly emitted from sources using
chloroform as a feedstock to produce another product
(fluorocarbons, EDC, pharmaceuticals, pulp and paper) or when
chlorine is used to disinfect water which then forms chloroform
(water treatment, POTW, cooling water treatment). Water
treatment accounts for 59 percent of the total chloroform
emissions in the five study areas. Pulp and paper manufacturing
and EDC production only occurs in the Baton Rouge study area and
together account for 18 percent of the total chloroform
emissions. Pharmaceutical manufacturing is a major contributor
to chloroform emissions in Philadelphia. Emissions from
pharmaceutical manufacturing, pulp and paper manufacturing, and
EDC production were estimated from plant specific data. To
estimate emissions from trichloroethylene (TCE) photodegradation,
cooling water treatment, and grain fumigation, the nationwide
emission estimates were apportioned by population (Mohin, 1984).
Drinking water treatment emissions were calculated using the
study areas' 1980 population figures, an average chloroform
concentration of 41 micrograms/liter (GCA, 1984), and water
consumption rates (Metcalf and Eddy, 1981). POTW emissions were
calculated using the methodology presented in the 35 County Study
(Versar, 1984). Baseline chloroform emissions are listed by
source category for each study area in Table II.5.
5. Chromium
The major sources of chromium emissions are chrome
plating, cooling towers, and refractory manufacturing. Chrome
plat~ng accounts for 34 percent of the chromium emissions in the
five study areas. Cooling tower emissions were estimated by
apportioning the nationwide emission estimate by population.
Radian provided local data to estimate emissions from
refractories. Chrome plating, municipal incineration, chemical
manufacturing, and iron and steel production emissions were
estimated based on the information provided in the document
"Study of Sources of Chromium, Nickel and Manganese Air
Emissions" (Radian, 1984a). Emission factors were used to
estimate emissions from coal and oil combustion (PEDCo, 1982) and
cement production (U.S. EPA, 19841). Baseline chromium emissions
are listed by source category for each study area in Table II.6.
6. Ethvlene Dichloride
. Major sources of EDC emissions are EDC production, and
miscellaneous uses, such as manufacturing of pesticides,
herbicides, and color film. EDC production accounts for 12
percent of the EDC emissions in the five study areas. EDC
production occurs only in the Baton Rouge study area and emission
estimates were made using plant specific data. EDC emissions
from miscellaneous uses were estimated by apportioning the
20
-------
Table 11.5
Chloroform
Baseline Emissions (TPY)
Los Baton
Source Category Phoenix Angeles Rouge Baltimore Philadelphia
Fluorocarbon Production 0.5
EDC Production* 6.6
Pharmaceutical
Manufacturing 22.8
Pulp & Paper 40.9
POlW' s 0.7 6.6 5.2 3.7
Grain Fumigation 0.2 0.1 0.1 0.2
Drinking Water Treatment 20.8 103.0 3.3 14.9 18.4
TCE ,Photodegradation 3.1 19.2 0.8 3.0 3.5
Cooling Water Treatment 1.9 12.0 0.5 1.8 2.2
Total 26.5 141.5 52.2 25.0 50.8
* Based on the emissions obtained fran the GCA chloroform assessnent document
for Ethyl Corporation and Formosa Plastics.
21
-------
Table II.6
Chraniun
Baseline Emissions (TPY)
Los Baton
Source Category Phoenix Angeles Roua:e Baltimore Philadelphia
Utility Boilers 0.1 0.7 0.5 0.5
Industrial Boilers 0.3 0.1 0.5 2.3
Iron and Steel
Manufacturing* 0.0 0.3
Refractory Manufacturing 12.5
Chemical Manufacturing 3.9** 0.1
Chrane Plating 15 5.1
Municipal Incineration 0.7 0.9
Waste Oil Burning 0.4 0.0 0.2 0.3 0.6
Cool ing TCMers 3.9 0.0 1.0 3.7 4.3
Cement Manufacturing 0.9
Total 4.4 16.9 1.3 22.4 13.8
* Emissions are primarily fran electric arc furnaces.
** Production of sodium chranate and sodiun dichranate.
22
-------
nationwide emission estimates by population. Gasoline marketing
emissions were calculated using county gasoline throughputs (U.S.
EPA, 1984k). Baseline ethylene dichloride emissions are listed
by source category for each study area in Table II.7.
7. Ethvlene Oxide
. Sources of .ethylene oxide are hospital sterilizers, food
sterilizers, and ethylene oxide production. Hospital sterilizers
emit 89 percent of the ethylene oxide in the five study areas.
Food sterilizers emit 9 percent of the total emissions but
emission estimates for this source type were only available for
the Los Angeles study area. Baseline ethylene oxide emissions
are listed by source category for each study area in Table II.8.
8. Formaldehvde
Major sources of formaldehyde emissions are mobile
sources and petroleum refining. Mobile sources em~t 63 percent
of the formaldehyde emissions in the five study areas. Petroieum
refining emits 3 percent of the emissions and is specific only to
the Los Angeles and Baton Rouge study areas. .
Formaldehyde emission estimates for all source
categories, except miscellaneous. sources, 'POTWs, and boilers,
were based on EPA emission factors (U.S. EPA, 1984e). Emissions
from boilers were calculated using the €mission factors presented
in the PEDCo paper (PEDCo, 1982). Miscellaneous source emissions
were estimated by apportioning the nationwide emission estimate
by population (SAI, 1982). The 35 County Study methodology was
used to calculate emissions from POTWs (Versar, 1984). Baseline
formaldehyde emissions are listed by source category for each
study area in Table II.9.
9. Gasoline Vapors
. By definition, gasoline vapors are organic emissions from
refineries and gasoline marketing. Major sources are service
stations (vehicle refueling and underground storage tanks), bulk
terminals (storage tanks and truck loading), and bulk plants.
Gasoline vapor emissions from bulk terminals and service stations
were estimated by apportioning the statewide emission estimates
by gasoline throughput (U.S. EPA, 1984k). For bulk plants the
statewide emission estimates were apportioned by gasoline
consumption and then multiplied by .the percentage of gasoline
cycled through each bulk plant (U.S. EPA, 1984k). Baseline
gasoline vapor emissions are listed by source category for each
study area in Table II.10.
10. Lead
Road vehicles dominate lead emissions in the five study
23
-------
Table II. 7
Ethylene Dichloride
Baseline Emissions (TPY)
Los Baton
Source Category Phoenix Angeles Rouge Baltimore Philadelphia
Gasoline Marketing 5.6 6.4 0.6 1.9 1.4
EDC Production* 53.3
Miscellaneous - EDC 39.2 242 9.9 37.4 43.8
Total 44.8 248.4 63.8 39.3 45.2
* This is the combined emission estimate for Ethyl Corporation and Formosa
Plastics. Emissions for each plant were obtained from MRI (1984).
24
-------
Table 11.8
Ethy lene Oxi de
Baseline Emissions (TPY)
Los Baton
Source Categorv Phoenix Angeles Rouge Baltimore Philadelphia
Sterilizers - Hospitals 7.8 325.0 2.8 9.7 11 .4
Sterilizers - Food 36.8
Ethylene Oxide Production 5.5 0.4
Total 7.8 367.3 2.8 9.7 11.8
25
-------
Source CategorY
Utility Boilers
Industrial Boilers
Residential-Commercial/
Inst. Boilers
Fabricated Metal Production
(misc. )
Petroleum Refining
Chemical Manufacturing
Phthalic Anhydride
Residential Wood Combustion
Motor Vehicles
Wood Production
Municipal Incineration
P01Ws
Residential-Commercial Gas
Furnaces
Railroad Locomotives
Miscellaneous-Formaldehyde
Total
Table 11.9
Formaldehyde
Basel ine Emissions (TPY)
Phoenix
0.4
9.6
2.6
137.0
Los
Angeles
557.4
6.4
14.8
115.0
419.5
43.5
. 178.4
1,197.65,182.1
262.8
5.1
9.7
70.2
19.6
19.7
125 .6
11.1
104.0
328.8
122.0
1,714.6 7,228.3
Baton
ROUlze
0.4
64.8
2.1
0.0
0.0
60.7
241 .8
4.4
19.2
5.0
398.4
Baltimore PhiladelDhia
1.8
8.1
2.3
6.8
56.9
79.6
96.2
0.3
64.8
115.9
629.5
35.2
543.1
374.4
75.4
1.1
756.0
837.2
13.3
27.8
22.4
60.4
18.8
21.9
1,418.1
2,431.1
Note: According to the SCAB report the total formaldehyde emissions fran
resin producing plants for this area is 0.4 TPY (0.36 MIS/yr). The only other
resin producing plant of concern is in Baltimore. The Bal timore plant has
very lcw VOC emissions reported in NEDS; applying a formaldehyde emission
factor to this number produced insignificant formaldehyde emissions.
26
-------
Table II. 1 0
Gasoline Vapors
Baseline Emissions (TPY)
Los Baton
Source Category Phoenix Angeles Rouge Baltimore Philadelphia
Bulk Terminals -
Truck Loading 1,937 219 101 332 248
Storage Tanks 370 296 30 4,421 72
Bul k Pl ants 1,800 212 236 332
Service Stations -
Underground Tanks 3,541 2,790 180 357 265
Vehicle Refueling 3,296 69 791 2,602 1,929
Petroleum Marketing Vessels 252
Gasoline Marketing 6,107
Total 10,944 9,733 1,314 7,948 2,846
27
-------
areas. However, small amounts of lead are emitted by fuel
combustion stationary sources and industrial processes. Emission
factors were used to estimate emissions from all source
categories. The PEDCo paper provided emission factors for coal
and oil combustion furnaces (PEDCo, 1982) and AP-42 was used as
the source of emission factors for all other source categories
(U.S. EPA, 1984a). Baseline lead emissions are listed by source
category for each study area in Table 11.11.
11. Nickel
Sources of nickel include stationary source fuel
combustion and industrial processes. Stationary source fuel
combustion accounts for 91 percent of the nickel emissions in the
five study areas. Emission factors were used to calculate nickel
emissions (U.S. EPA, 1984f). Baseline nickel emissions are
listed by source category for each study area in Table 11.12.
12. Perchloroethvlene (~
Major sources of PCE emissions are dry cleaning and
cleaning and degreasing operations. Dry cleaning emits 52
percent of the PCE in the five study areas. Dry cleaning,
solvent use, chemical manufacturing, and oil combustion emissions
were based on local data. POTW emissions were calculated using
the methodology presented in the appendix of the 35 County Study
(Versar, 1984). PCE emissions from cleaning and degreasing
operations were calculated by apportioning the nationwide
emission estimate by county populations. Baseline PCE emissions
are listed by source category for each study area in Table 11.13.
13. Products of Incomplete Combustion (tl.Cl
Sources of PIC emissions are wood smoke, road vehicles,
and residential coal combustion. These emissions were calculated
using the B(a)P emission factors presented in the 35 County Study
(Versar, 1984). It should be noted that B(a)P was used as a
surrogate for PIC emissions. Baseline B(a)P emissions are listed
by source category for each study area in Table 11.14.
14. Trichloroethvlene (~
Cold cleaners emit 68 percent of the TCE emissions in the
five study area~. Small amounts of TCE are emitted by
degreasers, POTWs, and chemical production. Emissions from cold
cleaners and degreasers were estimated by apportioning the
nationwide emission estimate (SAI, 1982) by the study a~ea's
population (U.S. DOC, 1983b). Local information was used to
calculate emissions from polyvinyl chloride (PVC) production,
solvent use, and chemical manufacturing. Baseline TCE emissions
are listed by source category for each study area in Table 11.15.
28
-------
Table II. 11
Lead
Basel ine Emissions (TPY)
Los Baton
Source CategorY Phoenix Angeles -
Rouge Baltimore Philadelphia
Utility Boilers 0.1 0.4 0.1 1.2 0.3
Industrial Boilers 0.0 0.2 0.4
By-product Coking .83.0
Iron and Steel 47.0 11.7 0.1
Primary Smelting:
Zinc .7.9
Lead 5.0
Copper 0.1
Secondary Slnel ting:
Lead 2.5 1,724. 1 9.9 46.8
Copper 180 .2
~tor V ehicl es 624.5 4,499.1 15t. 1 497.6 372.7
Municpial Incineration 0.0 37.4 75.6
Waste Oil Burning 105.5 470.8 56.9 69.1 151. 1
Cement Manufacturing 3.4
Glass Manufacturing 292.0
Total 732.6 7,230.0 224.0 700.2 647.0
29
-------
Table II. 12
Nickel
Baseline Emissions (TPY)
Los Baton
Source Category Phoenix Angeles Rouge Baltimore PhiladelDhia
Utility Boilers 1.0 7.4 2.2 5.1
Industrial Boilers 4.5 0.5 0.8 1.4
Iron and Steel 0.0
Refinery-Process Heaters 0.2 0.1
Residential Wood Combustion 0.3 0.3 0.2 0.2 0.1
Cement Manufacturing 0.0
Waste Oil Burning 0.1 0.5 0.1 0.1 0.2
Chemical Manufacturing 0.1
Total 1.4 12.9 0.9 3.3 6.9
30
-------
Table I1.13
Perchloroethylene
Baseline Emissions (TPY)
Los Baton
Source Category Phoenix Angeles Rouge Baltimore Philadelphia
Dry Cleaning 500.0 6,025 290.0 692.5 1,164.2
Solvent Use 43.9 3.0 9.2 14.0 68.5
Vapor Degreaser-
Solvent Use 7,668 8.0
Fabricated Metals
Production 1.7
Polyvinyl Chloride
Production 0.0
P01W 14.5 96.8 16.7 21.8
Waste Oil Burning 0.0
Chemical Manufacturing 0.1
Pulp and Paper 0.3
Total 558.4 13,792.8 299.2 723.2 1 ,264.6
31
-------
Table II. 14
Products of Incomplete Combustion (PIC)
Benzo(a )pyrene
Baseline Emissions (TPY)
Los Baton
Source CategorY Phoenix Angeles Rouge Baltimore Philadelphia
Residential Wood Combustion 0.5 0.4 0.1 0.4 0.1
Motor Vehicles 0.6 2.7 0.1 0.3 0.3
Residential Coal Combustion 0.1
Total 1.1 3.1 0.2 0.7 0.5
32
-------
Table II. 15
Trichloroethylene
Basel ine Emissions (TPY)
Los Baton
Source Category Phoenix Angeles Rouge Bal timore Philadelphia
Solvent Use 20.8 19.0
Open Top Vapor Degreasers 25.1 741.0 6.4 24.0 34.4
Cold Cleaners 650.0 0.0 166.0 620.9 727 .0
Conveyorized Vapor 3.4 0.0 0.9 3.2 3.8
Degreasers
POlyvinyl Chloride
Production 0.1 0.0
Chemical Manufacturing - 5.0
Adhesives
P01W's 12.1 87.5 11.5 9.3
Total 690.6 854.4 173.3 659.6 793.5
33
-------
15. Vinvl Chloride
Sources of vinyl chloride include chemical manufacturing
and POTWs. POTW emissions constitute 99 percent of the vinyl
chloride emissions in the five study areas. POTW emissions were
calculated using the methodology presented in the appendix of the
35 County Study (Versar, 1984). Chemical manufacturing emissions
were based on local data. Regulations specific to vinyl chloride
in the South Coast Air Basin have reduced emissions from
traditional sources there to negligible levels. Baseline vinyl
chloride emissions are listed by source category for each study
area in Table II.16.
34
-------
Table II. 16
Vinyl Chloride
Baseline Emissions
-------
III CONTROL TECHNIQUES EVALUATION
A. DEVELOPMENT OF EMISSION CONTROL DATA BASE
This phase was integral to the overall objective of this
project because it characterized three important elements:
. the degree to which existing (1980 baseline) emissions
are controlled and their associated costs of control,
. the degree to which both existing and new sources will be
controlled in the projection year (1995) and as a result
of new regulatory requirements imposed between 1980 and
1995, and
. the maximum level of control potentially achievable for
each source of toxic emissions.
Each source type ,emitting VOC or PM emissions was
included in a data base established by extracting three important
pieces of information from documents that evaluate emission
control techniques for specific source categories. The three
items are listed below:
. specific emission control devices and/or practices
applicable to each source type,
. the degree of control (percentage reduction from
uncontrolled emissions) provided by each technique, and
. the absolute cost effectiveness of each control technique
expressed in $/ton of PM or VOC reduced from uncontrolled
levels.
Because emission control cost effectiveness values are
very sensitive to facility size and whether the controls are
applied to new or existing facilities, as many as four sets of
cost-of-control data were established for each source category.
These four sets are (1) cost of control for new small facilities,
(2) cost of control for existing small facilities, (3) cost of
control for new large facilities, and (4) cost of control for
existing large facilities.
Most of the data for retrofit of controls to existing
facilities were derived from Control Techniques Documents and
Ba6kground Information Documents. New Source Performance
Standards (NSPS) provided data for new sources. Background
documents supporting NESHAPs provided data for both new and
existing facilities. All of these documents generally presented
control techniques data by various levels of stringency and for a
range of facility sizes.
36
-------
To match the control techniques data base, each
individual facility in the baseline emission inventory was
categorized as either "large" or "small" depending on whether its
size or processing capacity was determined to be larger or
smaller than the median size of all facilities in the same source
category. In other words, a median size was identified for each
SCC in the emission inventory. The size of each individual.
facility was then compared to the median value and the facility
was categorized accordingly.
For many source types, references for the control
techniques and associated costs could not be found. To account
for these facilities, generic techniques and control cost values
were established based on data contained in documents that
generically evaluate control techniques for vac and PM emissions.
For PM emissions, two documents were used (U.S. EPA, 1982b,
1982e).
For vac emissions, generic emission control cost
functions were developed from information from two other
documents (U.S. EPA, 1980a, 1980b). Tables III.1 and III.2 show
the generic cost effectiveness values developed from these
documents for vac and PM emissions, respectively. In Table
III.1, for vac emissions, cost effectiveness (CE) values are
. based on whether controls are applied eith~r to fugitive or stack
emissions and if the controls (e.g., incineration, carbon.
absorption), can yield credits either for product recovery or
heat recovery from incineration. Control costs increase if the
vac emissions must be scrubbed before control to remove sulfur
compounds or other materials that may "poison" catalysts or
activated carbon. .
. To calculate the CE values shown in Table III.2, it was
assumed that all PM emissions from each facility type could be
captured and controlled by a mechanical collector with a fabric
filter. No consideration was given to whether the collected PM
has product recovery value.
Appendix A (Radian) shows the data base of control levels
and associated cost effectiveness values applicable to each
facility type in the emissions inventory. As discussed in the
introduction of this report, a basic assumption used in this
project was that species of the criteria pollutants are
controlled to the same degree as the criteria pollutant itself.
Al though this is not universally true, the chemical and physical
characteristics of each emission stream would have to be
evaluated to do otherwise and differentiate among them. Thus,
the values indicating the degree of control of PM or vac
emissions were used to also apply to each toxic species of PM and
vac. The associated cost effectiveness values shown in Appendix
A (Radian) are expressed in 1984 dollars.
37
-------
Table 1:;:1.1
Ger.eric Cost Effectiveness ($/ton Pollutant ~educed)
Values for New' S~urce VOC Emissions (1984$)
arge
Fugitive Stack
Product No Product Scrubber Product No Product Scrubber
Recovery Recovery Required Recovery Recovery Required
113 478 720
24 720
48 574 720
197 552
-168
-120 250 528
Facility She ma
Type of Emission: Fugitive Stack
Type of Control Device: Product No Product Scrubber Product No Product Scrubber
Recovery Recovery Required Recovery Recovery Required
VOC Emission Control Cost Effectiveness $ ton
Control
Level (S Control)
72 278 607 864
w 75 384
co 79 408 698 888
90 262 624
98.6 -96
99 -48 312 600
.
For costs of retrofit to existing emission sources, multiply values In table by 1.3.
Hote: negative values represent a cost savings.
,
-------
Table III.2
Generic Cost Effectiveness
($/ton Pollutant Removed) Values for
New* Source PM Emissions (1984 $)
---------------------------------------------------------
-----------------------------------------------------------------
Control Level
(% Control)
~Q~~_£I£~~~j~~n~~~_Qf_~_Emj~~j~n~-~QDtI~l_l!LtQnl
Small Facility Large Facility
-----------------------------------------------------------------
98
99
353
415
139
163
99.6
581
228
Note:
These values are based on uSe of mechanical collectors and
fabric filters applied to stack emissions.
* Forcosts1of retrofit to existing PM emission sources, ~ultiply
values by 1.3.
-----------------------------------------------------------------
39
-------
B. USE OF CONTROL TECHNIQUES DATA BASE
1. General Use
The information input to this data base was used to
estimate the control costs for each source type under each
regulatory scenario examined, including the 1980 baseline
emissions case. To determine the appropriate cost effectiveness
value, the degree of control required of each source type was
first determined. For the base case (1980), the existing degree
of control for stationary sources is listed in the NAPAP emission
inventory files. Dividing these values into the annual PM and
VOC emissions reported for each source type yielded the .
"uncontrolled" emission rates. The difference between controlled
and uncontrolled emissions, expressed in tons/year, was
multiplied by the cost effectiveness value corresponding to the
level of control required for the source type to estimate an
annualized ($/yr) cost of control.
The calculation of control costs for the base period
(1980) used cost effectiveness values for control applied to
existing sources (retrofit). Although some of these controls are
not retrofits, it was impossible to segregate from the base
inventory those emissions that reflect controls that were applied
at the time the emission producing facility was constructed.
Thus, the emission control costs for the base period (1980) may
be somewhat overstated because the cost effectiveness values
associated with retrofit are greater, in many cases, th~n the
same control technique applied .to a new source. However, control
costs estimated for future years considered the appropriate cost
effectiveness value applicable to either new sources or retrofit
of controls to existing sources.
2. Mobile Sources
For mobile sources, cost effectiveness values for control
of PM and VOC emissions were calculated based on annual per
vehicle costs of the Federal Motor Vehicle Control Program
(FMVCP), Inspection and Maintenance Programs, and the Lead
Phase-Down Program. Equipment costs for control of VOC emissions
for light duty gas vehicles subject to FMVCP were based on
information derived from Lindgren (1978). The component costs
were adjusted to 1984 prices and were estimated to be
$601.42/vehicle. Assuming a 10-year life for each vehicle, a 10
percent discount rate and no operat~ng cost penalty yielded an
annualized cost of $97.88 per year per vehicle. For light-duty
gas vehicles, an annual usage rate of 10,756 miles traveled was
assumed. The difference in hydrocarbon emission factors (9.05 gm
HC/VMT, uncontrolled and 1.64 gm/VMT, controlled) was used to
calculate control levels and the cost effectiveness of VOC
emission reductions associated with the FMVCP.
40
-------
Because each study area has implemented an Inspection and
Maintenance (I/M) program for light duty gasoline vehicles, the
incremental VOC emission reductions and associated costs of
control were calculated considering the type and degree of
stringency of the I/M programs for each area. Inspection costs
for all vehicles (those subject to FMVCP in addition to those of
earlier model years) were assumed to be $10 per year vehicle.
Failure rates for Los Angeles, Baltimore, Philadelphia, and
Phoenix were assumed to be 20 percent. The Baton Rouge failure
rate was assumed to be 3 percent. The average combined HC
emission factor for FMCVP vehicles in Baltimore, Philadelphia,
and Phoenix due to I/M programs was estimated to be 1 .4102
gm/VMT. Because of the stringent I/M program in Los Angeles, an
HC emission factor of 1 .3172 gm/VMT was used. The Baton Rouge
anti-tampering I/M program was estimated to produce an aggregate
VOC emission factor of 1 .547 gm/VMT. Repair costs per "failed"
vehicle for I/M assumed an annual average cost of $50 ($150 for
Baton Rouge).
Control of PM emissions for gasoline vehicles was assumed
to result entirely from reduction in gasoline lead content to 0.1
gm/gallon. Cost and PM emission estimates were extracted from-
U.S. EPA (1984m). As presented therein, annual costs of lead
phasedown average $7.50 per vehicle per year and the PM emission
factor is reduced from 0.0886 gm/VMT (uncontrolled) to 0.0098
gm/VMT. Another hypothetical level of PM control was estimated
based on a completely lead-free gasoline fleet which reduced the
PM emission factor to 0.005 gm/VMT. The cost of this option was
assumed to be $22.50 per vehicle per year.
Similarly, cost and emission control estimates were made
for heavy-duty gas vehicles and diesels. Data were obtained from
U.S. EPA (1979b) and Automotive Engineering (1984) to calculate
cost effectiveness values for these classes of mobile sources.
For each category of mobile sources, emission projections
for each scenario were made using the MOBILE3 computer model.
MOBILE3 predicts the rates of vehicle replacement and vehicle
miles traveled by each model year of vehicle for each area.
Thus, using the emission factors and cost data presented above,
cost effectiveness values for each category were calculated and
are presented in Table III.3. The data in this table were used
to estimate the cost of each alternative control program
(scenario) according to the cost effectiveness.of the selected
program and tons of PM and/or VOC r~duced by the program. In
other words, MOBILE3 was exercised twice for each scenario: once
using the uncontrolled emission factors and the second time using
the emission factor corresponding to the emission control program
examined. Multiplying the difference in emissions (uncontrolled
minus controlled) by the appropriate cost effectiveness value
produced the estimated control program cost for each area.
ql
-------
Table III.3
Cost Effectiveness and Erni~sion Reduction Values
for Control of Mobile ~ource PM and VOC E~issions
\.
J
Vehicle Class
Emis s 1 on
. Control Program
Applicability
PM Control
Cost Effectiveness
~ Reductlon* (S/Ton Reduced). S Reduction-
VOC Control
Cost Effectiveness
(S/Ton Reduced)-
Light Duty Gas Vehicles
FMVCP
11M
11M
11M
Lead phase down
Lead phase out
All areas
Phoenix ,Ba 1 to ,Phlla.
Baton Rouge
Los Angeles
All areas
All areas
88.91
94.4
8030
22706
81.88
84..4
82.9
85.45
1114
1302
1207
1287
[[[
.r: Light Duty Gas Trucks FMVCP All areas 78.38 1068
I\J 11M Phoenlx,Balt.,PhfJa. 82.36 1224
11M Baton Rouge 79.41 1158
11M Los Angeles 83.37 " 1209
Lead phase down All areas 81.49 6479 .
Lead phase out All areas 94.36 18084
[[[
Heavy Duty Gas Vehicles
Federal Standards(partlal)
Federal StandardS(full)
Lead phase down
. Lead phase out
All areas
All areas
All areas
All areas
88.69
94.36
14241
40169
43.3
86.5
213
299
-------
IV EMISSIONS AND COST PROJECTIONS
A. NEW SOURCE GROWTH RATES
Growth in new source emissions was estimated for the
1980-1995 period using two-digit Standard Industrial
Classification (SIC) level of detail for each industry.
Population was used to estimate growth for some non-industrial
source types. Industry growth rates are from OBERS, published by
the U.S. Department of Commerce Bureau of Economic Analysis, and
represent growth in industry earnings. Population projections
are from the same source. Because Standard Metropolitian
Statistical Area (SMSA) level industry earnings projections are
only available at the one-digit SIC level, the method used to
provide two-digit SIC growth rates for this project was to
allocate state level two-digit growth estimates to each SMSA
using that SMSA's share of growth at the two-digit SIC level. It
is Pechan's belief that this is a reasonable compromise between
simply providing one-digit SIC growth rates for each SMSA and
assuming state-wide growth rates are representative of any area
within that state. Table IV.1 shows the growth factors by
industry for each study area.
B. ESTIMATES OF FUTURE UTILITY EMISSIONS
Because utilities are generally an important source type
in any air pollution study and data are available on planned
utility installations through 1995, we performed a separate
growth analysis for this source category. We investigated
whether any power plants burning coal, oil, or gas are planned
for the five study areas for the 1980-1995 period. The
unexpected finding of this analysis was that no significant
utility emitters of toxic air pollutants are planned for the 15
year time horizon. The only planned facility for any of the
study counties is a .50 megawatt gas turbine to be located in Gila
Bend, Arizona (Maricopa County). Planned completion of this unit
is May 1992, but because it will burn natural gas, it will not be
a significant emitter of toxics.
Therefore, unless otherwise reduced via one of the
control scenarios, utility emissions are assumed to be the same
in 1995 as they are in 1980.
C. ESTIMATED EFFECTS OF ALTERNATIVE CONTROL PROGRAMS
Because of the large amount of data to be processed and
the number of alternative scenarios to be evaluated, a computer
program was developed in order to perform many iterations in a
short period of time. Thus, Radian developed the Regulatory
43
-------
Ta bl e IV. 1
Population and Economic Growth Rates
State Two-digit SIC Allocated to Selected SMSAs
Los Baton
SIC Category Phoenix Angeles Rouge Baltimore Philadelphia
Population 1.34 1.14 1.22 1.04 1. 0 1
01-07 Agricul ture 1.18 0.95 1.23 1.25 1.02
10 Me tal Mining 1.17 NA NA NA 0.90
1,1-1 2 Coal Mining NA NA 2.84 1.55 1.69
13 Oil and Gas NA 1.01 1.02 5.71 1.22
1 4 Nonmetal Mi ni ng 1.88 1.19 NA 0.91 1.03
.l::" 15-17 Construction 1.54 1.35 1.35 1. 31 1.36
.l::" 20 Food 1.48 1.18 1.28 1.14 1 .09
21 Tobacco NA 0.97 0.48 NA 0.68
22 Textiles 1.07 1. 75 NA NA 0.84
23 Apparel 1.62 1.55 1. 71 1.10 0.93
24 Lumber 1. 72 1.21 1.33 NA 1. 13
25 Furniture 1.94 1.39 1. 13 NA ' 1 .21
26 Paper 2.00 1.47 1.50 1. 25 1.25
27 Printing 1.96 1 .52' 1.66 1.28 1.26
28 Chemical s 2.15 1.65 2.23 1.30 1.35
29 Petroleum Refining 1.92 1. 18 1.60 1.27 1.06
30 Rubber and Plastics 2.50 1. 70 NA 1.12 1.57
31 Leather 1.52 1.16 1.09 NA 0.73
32 Stone, Cl ay, and Glass 2.22 1.43 1.69 1.33 1.24
33 Pr imary Me tal s 1.62 1.32 2.90 1.19 1.15
34 Fabricated Metals 1. 75 1.48 1.88 1.15 1.16
35 Non-electrical Machinery 1.95 1. 74 2.66 1.35 1.20
36 Electrical Machinery 1.87 1.44 2.99 1.43 1. 21
37 Transportation 1.50 1.07 1.90 1.14 1. 44
38 Instruments 1.52 1.65 2.38 1. 21 1 .19
39 Misc. Manufacturing 1.46 1.36 1.52 0.94 1.04
72 Serv ice s 1. 91 1.63 1.87 1.47 1. 54
-------
Impact Model (RIM); its structure and function are discussed
below.
1. DeveloDment of Analvtical Framework
Shown in Figure IV.1 is a schematic of the Regulatory
Impact Model (RIM) that was used to estimate future emissions and
costs of emission control ~or stationary sources of PM and VOC
emissions. Costs for mobil€ sources were developed separately
and are discussed later in this section.
As shown in Figure IV.1, there are two functional
components of RIM:
. the emissions projection module and
. the control cost module.
Operation of RIM starts with the baseline emissions
inventory which, for each source type, contains information
relating to annual emissions and the level of control in place on
each source type. From this, the uncontrolled emission rates can
be determined.
In order to project changes in the baseline emissions to
any future year, three pieces of informafion, specific to each
source type, must be established:
(1) the rate at which old equipment will be replaced with
new, less polluting equipment,
(2) the rate at which the industry (or emissions source
category) is expected to experience growth in a
geographic region, and .
(3) the constraints that existing and future environmental
regulations impose on sources to reduce emissions from
uncontrolled levels.
Accordingly, three files were established, each of which
operate on the uncontrolled PM and VOC emission levels of each
source category in the baseline inventory. These are discussed
below.
a. Equipment Replacement Rate File
. This file was established to estimate emission changes
that occur when old industrial equipment is replaced with new,
lower polluting equipment. Replacement rates of old equipment
and its associated uncontrolled emissions were assumed to be two
times (2x) the Internal Revenue Service depreciation schedule for
industrial equipment (IRS, 1983). The RIM software removes the
45
-------
DATA ""o.T
lUll..
1_8»I8t
-nlfT08T
.t="
0"\
CONnO'-
COlt
.. OOC.H.I
COIITIO'-
con.
Figure IV.l
Schematic of Regulatory Impact Model (SI~)
ItIW CA'ACITT MOWTN ~
AltO l.,ucnUT UTi.
l"tt'ION CON,uun,
~
"W MOWTN
I'" II lOtI.
"OUUTtON'
A"lICUU
TO All
I"" ,tOIl .
II'"ACIY...T
A-.o
MeW ICXMCI
OIOWTN
(Nln, I")
I.,UCI"..,T
1l1li11 IOt8 I
.lunltO
I""~
(UCT, I",
NIIHU',
.,"")
t
IIIIttllON "IOJrCTION MODUU
"W .OUIICI
MOWTN IYttltONI
II'UCI"..,T
IM"'tONl
IlIn...o
1"'"1<»11
8ftW lou.CI COIIT"O'-
con ,UItC T IOtC8
"IUO'IT CONT"OL
COlT 'UMCTtOQ
OUT",T
HOJtCTlD CONTIIIOl coni
nw lou.CI
caOWTN
(UCT, UII)
JIYIIIO
...--.-.
UNCOMTIOUID
IM'"I088,
ANO
-------
COIlTlOl
unL
t
O(JTMlT
"oJCcno
&M.II.08l,
-------
uncontrolled emissions associated with replacements from the
"existing emissions" file and places them in a "replacement
emissions" file as shown in the example below:
ExamDle of Replacement Rate Calculation
Baseline Uncontrolled PM Emissions for SCC 10200401
(Industrial Boiler-oil fired) = 100 TPY
IRS depreciation life = 20 years
Replacement life = 2 x 20 = 40 years
Annual Emissions Replacement Rate = 1/40 x 100 = 2.5%
per year
Replacement Uncontrolled PM emissions after 1st year =
100(.025) = 2.5 TPY
Existing Uncontrolled PM emissions after 1st year =
100 - 2.5 = 97.5 TPY
This procedure is repeated for each year of the
projection period and simultaneously depletes the existing
emissions file by the amount added to the replacement emission
file. This process accounts for the influence of NSPS and other
state and local regulations that may apply to new equipment at
any time during the projection period. The uncontrolled
emissions are reduced by the degree of control specified by the
applicable regulation for each year that the regulation is in
effect. In other words, a regulation that is expected to be
effective during the projection period will affect only those
uncontrolled emissions associatedwith replacements occurring
after the regulation goes into effect.
b. New Source Growth Rates
Table IV.1 presented the industry specific economic
growth factors covering the period 1980 to 1995 for each study
region. These values are input to a "growth rate" file after
converting them to an annual growth rate. For each year of the
projection period, new source uncontrolled emissions are
calculated by multiplying the baseline emissions (uncontrolled)
by the annual growth factor for each emissions source category-.
This procedure enables RIM to estimate the effects of
environmental regulations that apply to emissions growth such as
the control technology requirements (BACT, LAER) associated with
Prevention of Significant Deterioration (PSD) and Nonattainment
Area New Source Review (NSR). Furthermore, this procedure
provides the flexibility to estimate effects of more stringent
NSR regulations that may take effect at some point during the
projection period (1980-1995).
47
-------
c. Constraint File
The constraint file simulates the effects of emission
regulations or any other situation that produces reductions in
emissions from uncontrolled levels. As shown in Figure IV.1,
there are three general categories of constraint applicability:
(1) Constraints that affect onlv new source growth.
Generally, these are case-by-case control technology
determinations required by EPA and state-specific New
Source Review Regulations - Best Available Control
Technology (BACT) and Low~st Achievable Emission Rate
(LAER). These types of requirements are generally the
most stringent. In the case of the wood smoke category,
one constraint simulated the effects of new wood stove
designs that are inherently lower polluting than the
design upon which emission factors are based. Thus, the
emerging design and its emissions performance level
inherently "constrains" uncontrolled emissions from wood
stoves.
(2) Constraints that affect new source growth and eauipment
replacement onlY. This category of constraints
simulated the effects of NSPS and other state-specific
regulations governing new equipment.
(3) Constraints that affect all emissions. This. category
represents regulations that require emissions control of
all sources, regardlesa of whether they are existing or
new. In many cases, except for National Emissions
Standards for Hazardous Air Pollutants (NESHAPs), these
constraints are the least stringent. This is not
necessarily true for Los Angeles, where many of the local
air pollution regulations that affect existing sources
are more stringent than NSPSs and even Federal NSR
requirements.
The effect of all constraints is expressed in terms of
the degree of emission reduction (e.g., control percentage) over
uncontrolled emission rates. Each constraint produces a
controlled emission level (tons/year PM or VaC) by multiplying
the uncontrolled emission level of an emission category
(existing, replacement, or new) by the control percentage of each
applicable constraint and subtracting this product from the
uncontrolled emission level. Because many different regulations
may affect an emissions category of a single source type, the RIM
software selects the most stringent one, defined as that which
produces the lowest emissions level during the year examined. At
the final year of the projection period, emission changes
resulting from new growth, equipment replacement, and regulatory
requirements are calculated for each source type by size (large
and small). Thus, for each source/size category, the
48
-------
uncontrolled and controlled emission levels are calculated and
the difference between the two is filed as "tons per year
controlled." From this, the average pollutant control level for
each source/size category is calculated. One boundary condition
on all constraint calculations is that the degree of control of
any emissions category can be no less that the existing level of
control reported for the sources in the base year (1980)
emissions inventory. In other words, even in the absence of
directly applicabl~ regulations, the degree of control for new
emissions will be no less than that for existing emissions.
The output of the emission projection module produces
four records for each source category of emissions for each
geographic area as shown in Table IV.2.
2. Development of Emission Constraints Data
Emission constraints (level of emissions control) in the
form of current and potential criteria pollutant regulatory
programs and NESHAPs were developed for all SCCs in the five
study areas. Constraints included all Federal and state
regulations that directly affect PM and VOC emissions during the
base year (1980) and future regulatory programs that might affect
sources through 1995.
As noted in the preceding di~cussion; control levels were
established for three separate categories of potential emissions
-- existing, replacement, and new source growth -- in both.
attainment and non-attainment areas. . Emission constraints were
identified for each category. Federal and state requirements
considered for existing sources included the following:
. Reasonably Available Control Technology (RACT) in
non-attainment areas,. .
. State Implementation Plans (SIPs) in attainment areas,
and
. NESHAPs.
Emission constraints covering replacement and new source
growth are shown in Table IV.3.
Principal information sources for the emissions
constraint data base were the Federal Register, Code of Federal
Regulations (CFR), Control Technologv Guidance (CTG) documents,
SIPs, and the Bureau of National Affairs (BNA) reporter. These
sources provided an initial set of measures to use in defining
baseline (existing) source control levels and controlled/
uncontrolled emissions in each urban area. To ascertain
what additional emission control programs were in the process
of being implemented, or under serious consideration for
49
-------
.;
Ta bl e IV. 2
Emission Projection Module Records
.B~gj..Qn: xxxxxx
-SJ;;L-='.;-._.x.x.x.x.x.x
Uncontrolled
.E.m.1~.:2.i.Q.D~illIl
1. Large facility-existing
2. Large facility-new
3. Small facility-existing
4. Small facility-new
50
Controlled
.E.mil.:2.i.Q.D~iI.fIl
Emissions
.fig.d'y,g~JLillIl
-------
T a bl e IV. 3
Emission Constraints
--~J21.,g.Q~.ID.e.D.t_.a.n.L~.rL.s.Q.Y.r.Q~_G.r.Q!l.t.b_-
o New Source Performance Standards (NSPS)
o SIP Regulations
o New Source Rev iew (NSR)
51
.N~.1L-S.Q.Y.r.QLG.r.Q.H.t.b
o BACT
o LAER
-------
future implementation (between 1980 and 1995), Radian contacted
state officials within each study area and EPA staff who have
responsibility for the pertinent criteria pollutant programs.
Areas of review in determining more stringent regulatory measures
included the following:
. SIP provisions for characteristics of specified sources
to be controlled,
. mobile source emission regulations (e.g., requirement for
on-board vac controls, IIM, anti-tampering, lead
phasedown, and anti-misfueling programs),
. NSR program policy, and
. PSD program options.
Local regulatory control and plans were considered where
appropriate (e.g., particularly as applicable to mobile source).
Emission constraints were also developed for new NESHAP
initiatives deemed likely to have an effect by 1995. These were
derived from analyses conducted by EPA's Pollutant Assessment
Branch (U.S. EPA, 1985).
As illustrated in Figure IV.2, regulatory
collected for emission sources was entered into a
format as the emissions constraint input to RIM.
included the following:
information
uniform coding
These da ta .
. the pollutants controlled;
. sources affected;
. applicability of the constraint to new, replacement, or
existing emissions;
. the general timing of the application of a specific
regulatory measure;
. degree of control required (percentage of control applied
to uncontrolled emissions);
degree of penetration (percentage of emissions not
exempted from a requirement); and
. the year the constraint takes effect.
For this study, all the above regulatory information was
compiled for each source subcategory by the size of facilities in
each. That is, SCCs were disaggregated into large and small
facility subcategories, and control levels were established for
52
-------
pollutant:
NZlme:
Geo. Appllc:
Appllc SCC( s):
Figure IV.2
Constraint Code Sheet for Input to RIM
AIR TOXICS
Er-II SS ION CONSTRA I NT COO I NG FORM
(80 char maximum)
New Pres?:
Seoment Aopl 'cab I I Ity
ExIst PrEs?:
Repl. Pres?:
ATTainment?:
Nonattn?:
Time Frame
Beg Year:
TargeT YeZlr:
ConstraInt
Ab s Contro I :
PeneTraTion:
CommenTS:
(240 char max)
(Y or N)
(Y or N)
(Y or N)
(Y or N)
(Y or N)
(Beg year when constraint takes effect)
(Year when conSTraint takes full effect)
(Pret of emissions affected)
53
-------
each. The rationale for this disaggregation is that state and
Federal regulations sometimes include size exemptions below which
facilities are exempt from regulation. Furthermore, control
costs can vary considerably with facility size and this procedure
allowed this factor to be taken into account.
Several methods and assumptions were key to development
of the baseline degree of control for all subcategories. First,
for direct control, all regulations that affect source emissions
were identified as a constraint. The absolute degree of control
required by each regulation was determined using a hierarchy of
methods, beginning with a prescribed minimum percentage reduction
as stated in the regulation, and moving in successive steps
through (1) implicit degree of control (acceptable equivalency as
a percentage reduction) as stated in. a Federal Register notice or
other reference, (2) calculated level of control (using
uncontrolled rates listed in AP-42 -- U.S. EPA, 1984a), to (3)
best engineering judgment. In addition, target years when a
constraint takes full effect were identified as occurring at a
specific year from 1980 to 1995. Regulations affecting "all"
(new, replace~ent, and existing) sources were assumed to have
full effect by 1995.
Secondly, the total amount of a subcategory's emissions
affected by a prescribed constraint was defined in terms of a
"penetration" fa6tor. This factor serves as a measure of the
percentage of emissions for a source actually controlled by a
regulatory requirement. For examplet all regulations that embody
an applicable size threshold were assumed to affect emissions
only at "large" facilities. In those cases, the penetration
factor was 100 percent. Similarly, some sources within a given
category maybe exempt from a.regulatory requirement, for one
reason or another, and would therefore yield a penetration factor
of less than 100 percent. Furthermore, some sources only
partially covered by a requirement in the base year were
scheduled for incremental increases in control levels over the
study time frame. This was the case especially for controls
defined in the pertinent SIPs.
Finally, as discussed in
growth rates were calculated for
(tons/year) based on a specified
emissions.
the preceding sections, annual
SCCs as a fixed value
percentage of baseline
Appendix B (Radian) provides a complete listing of
emission constraints which reflect the required degree of
reduction over uncontrolled emissions for each source in the
study areas. The controls identified there were assumed to be in
place as a result of state and/or Federal requirements in effect
in the base year and projected to be in place by 1995. The next
section described Radian's approach to estimating future (1995)
emissions for six scenarios:
5~
-------
Scenario 1 is representative of current and likely
regulations in effect by 1995, and
five hypothetical scenarios increasing stringency over
Scenario 1.
After establishing the source subcategory baseline
control level, RIM was used to project both total subcategory
emissions and aggregate degree of control for regulatory
requirements likely to occur under current criteria and NESHAP
pollutant programs and for more stringent regulatory programs
hypothesized for the five study areas. The two principal
components of this analysis were (1) development of future
regulatory scenarios in order to quantify the discrete and
incremental effects of alternative strategies and (2) prediction
of controlled PM and vac emissions in comparison with
uncontrolled levels.
To estimate emission changes to 1995, it was necessary to
consider the following:
. facility/equipment replacements,
. new capacity growth, and
. increased (more stringent) regulatory requirements that
may reasonably be in place by 1995 and not reflected in
baseline emissions.
Separate adjustments for facility replacements and
capacity growth were made to the baseline calculations of
controlled emissions. As previously discussed, equipment
replacement was dealt with by considering both equipment life and
the range in regulatory requirements for the source type (taken
from information developed for the emissions constraint file).
For replacement rate, it was assumed that replacement will occur
at a rate commensurate with equipment life. The equipment life
factor was assumed to be two times the IRS depreciation life.
For example, the IRS depreciation life for a source is 10 years.
For purposes of this study, if it is assumed that equipment will
be replaced every 20 years, it would follow that 5 percent
(1/20th) of the uncontrolled emissions will be replaced every
year and those uncontrolled emissions will be subject to the
emission constraints applicable to replacement facilities (e.g.,
NSPS). .
The relationship of the above to changing regulatory
requirements can best be explained by illustration. If an NSPS
exists (or will exist) for a source type, the replacement unit
may be required to control emissions to the NSPS level or to a
SIP level in attainment areas or to a RAC! level in
55
-------
non-attainment areas, whichever is most stringent. Therefore, a.
5 percent per year annual turnover (as discussed above) may move
these sources to a control level of increased stringency. .
Because RIM uses uncontrolled emissions as a starting
point for estimating reductions, the only factor that affects
uncontrolled emissions is new growth. For the capacity growth,
new sources were assumed to be subject to either BACT, as .
required by PSD regulations applicable in attainment areas, or
LAER as required for new sources in non-attainment areas. This
growth in "new" emissions is thereby projected at control levels
reflective of greater stringency than existing emissions. To
determine BACT and LAER control levels, data were obtained from
EPA's BACT/LAER clearinghouse and Radian's New Source Review data
base of PSD permits. .
For prospective regulatory requirements, scenarios
(discussed below) were developed to identify both affected
facility populations and the degree of required control to be
applied to the population (or subpopulation) of sources in a
subcategory. Once emission constraints data were coded to embody
these requirements, RIM was used to calculate the percentage
change in controlled PM and VOC emissions from the :1980 baseline
and also the increased aggregate degree of control associated
with each hypothesized scenario. In sum, the emission reductions
result either ffom requiring reductions on sources that otherwise
would not be affected or from a greater degree of control for a
currently regulated source.
A total of six different regulatory scenarios were
constructed to project 1995 PM and VOC emissions by major source
category and study area. These six scenarios are described below
and in Table IV.4.
Scenarios 1a through 2c reflect the effort to "optimize"
on specific regulatory requirements as they apply to the study's
source subcategories (or parts thereof). More specifically,
Scenario 2a quantifies the impact of focusing on more stringent
New Source Review control requirements; Scenario 2b adds the
incremental effect of tighter controls on mobile sources (by
applying California emission constraints to the other study
areas); and Scenario 2c quantifies the effects of requiring
stringent NSPS control on all replacement emissions. For each
source category, the stringent control requirement was defined as
the control technique that reduces .PM or VOC emissions to the
greatest degree at a cost no greater than $4,000/ton controlled.
For each source category, the control techniques were identified
iIT the control techniques cost files and superimRosed on the
emissions affected in each scenario.
Finally, Scenario 3 represents an "upper limit" or
maximum potential benefits (measured in tons of PM and VOC
56
-------
Scenario 1:
Scenario 1a:
-Scenario 2a:
Scenario2b:
Scenario 2c:
Scenario 3:
Table IV.4
Description of Regulatory Scenarios
Emissions projections for 1995 under existing and
expected criteria and NESHAP pollutants regulatory
programs. Results of this scenario are considered
representative of the 1995 emissions plcture if
only the current regulatory agenda is accomplished.
Same as Scenario 1, plus the effects of new NESHAP
initiatives. This scenario might result if EPA
focuses control of toxic air emissions on Section
112 of the Clean Air Act resulting in significantly
more NESHAPs.
Same as Scenario 1a, with the addition of the most
stringent (reasonable) controls on new capacity
emissions. The incremental effect of this scenario
may be described as requiring very stringent BACT
on all new sources of air toxics.
Same as Scenario 2a, with the addition of the most
stringent controls on road vehicle emissions. This
scenario imposes the control of mobile source
emissions required in Los Angeles to all study
areas.
Same as Scenario 2b, with the addition of the most
stringent controls (reasonable) on replacement
emissions. The scenario extends stringent BACT to
replacement sources of toxic air emissions.
Same as Scenario 2c, with most stringent
(reasonable) control on all (new, replacement,
existing, retrofit) emissions. This may best
represent a requirement for stringent BACT on all
air toxics sources.
57
-------
controlled in 1995) of extending stringent control requirements
to existing sources. The results of Scenario 3 identify the
bounds to which the most stringent requirements under existing
criteria and NESHAP programs might reduce toxic PM and vac
emissions in the five urban areas.
3. Effects of Current and Future Regulatorv Requirements
on Toxic PM and vac Emissions and Control Costs
Figures IV.3 through IV.7 show for each study area the
total uncontrolled and controlled toxic PM emissions for each
scenario. The values shown reflect the PM emissions of those
sources that emit at least one of the five toxic PM species
involved in this study. Scenario 0 represents the baseline
(1980) while all other scenarios represent the emissions picture
expected in 1995 under the conditions of each scenario. Figure
IV.8 shows the totals for the five study areas.
Figures IV.9 through IV.14 show the same information
except for toxic vac emissions. . Again, the 10 toxic vac species
examined in this study constitute only a portion of the total vac
emissions. In other words, each source contributing to the total
vac emissions shown on these figures emits some amount of the
toxic pollutants involved in this study.
rable IV.5 shows for each study area, the total amount of
PM emissions reduced by application of control techniques
required by the various regulatory programs associated with each
scenario. As shown, the data are subdivided into mobile and
stationary sources. The total PM emission reductions shown for.
each area are the differences between the uncontrolled and
controlled PM emissions shown in Figures IV.3 through IV.8. The
cost data shown represent the total estimated costs of
controlling the indicated amount of PM emissions. Except for
road vehicles, the cost data are presented both in absolute terms
and relative to baseline (1980) costs. For road vehicles, all
costs are presented relative to. the baseline (1980) control
costs. This is because of the difficulty in identifying the
costs associated with gasoline lead content requirements in
effect in 1980. Therefore, the 1980 PM control costs for road
vehicles are assigned a value of zero (0) and 1995 control costs
are presented as an incremental cost over baseline.
Table IV.6 shows the same data as Table IV.5 except for
toxic vac emissions. The total VaC.emissions reduction
represents the difference in the controlled and uncontrolled
emissions for each scenario as shown in Figures IV.9 through
IV;14. Baseline (1980) costs for both stationary and mobile
sources are presented.
Table IV.7 summarizes Table IV.5 and IV.6 and presents
incremental costs of toxic PM and vac emissions control
58
-------
Figure IV-3. Tota1 Uncontrol1ed and Control1ed Toxic PM Em1ssions by
Scenar10.
Geographic Area: Phoenix
20000
18100 18100 1&100 1&100 1&100 18100
1&000
Vl 16000
\.0
14000 12700 12700
12500 12500
12000
Em Isslons 10000
( tonslyr) .
8000
6000 .
..000
2000
0
o (Ba..: I~O) I (1995) la(I995) 2a(\995) 2b(I995) 2c(I995) 3( 1995)
ROQu1alory Scenario
-------
Figure IV-4. Tota1 Uncontrol1ed and Controlled Toxic PM Em1sstons by
Scenart o.
Geographic Area: Los Ange1es
900000 470~O e70000 e70~O e70~O "70000 e70bJO
eooooo
0\ 715700
0 700000
600000
Emissions 500000
(tons/yr) -400000
300000
. 209&00
200000
I 00000
0
O(aDe&; 1~0) 1 (1995) 10(1995) ~. (1995) 2b(I99S) 2c(I99S) 3( 1995)
Regu1otory Scenorio
-------
F1 gure I V-So
900000
aooooo
0\
.....
700000
600000
Emissions 500000
(tons/yr) <400000
300000
200000
I 00000
Total Uncontro11ed and Control1ed ToxIc PM Emlss10ns by
Scenar10.
GeographIc Area: Baton Rouge
o
a2S1OO ~2SiOO "2SiOO ~2SiOO 825400 a250400
360400.
1~00 20500 20300 25000 25000 25000 23100
o (Bo8t: 1~0)
I (1995)
10(1995) 20(1995) 21>(1995)
Revu1atory Scenario
2c(I99S)
3( 1995)
-------
FIgure I V-5.
Tota1 Uncontrol1ed and Contro11ed Toxic PM Emtsstons by
- ScenarIo.
Geographtc Area: Ba1ttmore
90000
80100 80100 eOIOO. eOIOO eOI00 80100
~oo
0'\ 70000 66500
f\)
60000
Em 1ss1ons 50000
( tonslyr) ,"0000
30000
20000
10000 e300
3~0
o
0 (Boat: lo.eO) 1( 1995) 10(1995) 21(1995) 2b(I99S) 2c (1995) 3( 1995)
RB9ulatory Scenario
-------
Figure IV-7. Tota1 Uncontro11ed and Control1ed Toxic PM Emissions by
Scenario.
Geographfc Area: Phl1ade1phfa
00000
53300 53300 53300 53300 53300 53300
50000 "0000
0\
W "0000
Em 1s:slons 30000
(tons/yr)
20000
10000 80:10 MOO 8400 8300 8300
0
O(Bau: 1080) 1 (1995) la(1095) 2a(I995) 2b(I99S) 2c(I99S) 3( 1995)
Regulatory Scenario
-------
F1gure IV-B.. Total Uncontro11ed and Contro11ed ToxIc PM Emlss10ns by
Scenar10.
GeographIc Area: A11 Areas
2000000
1&-47500 16'17500 1 &<17500 16'17500 1 &<17500 1&-47500
1 &00000
1600000
0\
J:::o
14QOOOO
1220800
1200000
Emissions 1000000
Ctons/yr)
aooooo
600000
..00000
253200 2~400 247&00 24(J4(l0 24(J4(lO 243000 23()3Q0
200000
0
0 (Belt: 1030) 1 (J 995) lo(J995) 2o(I~5) 2b(I995) 2c (J 995) 3( 1995)
RBtJulatory Scenario
-------
FIgure lV-g. Tota1 Uncontrol1ed and Contro11ed Toxic VOC EmissIons by
Scenario.
Geographic Area: PhoenIx
250000
223000 223600 223000 223600 223000 223000
0\ 200000
V1 1&1000
150000
Emissions
(tons/yr) 110050
100000
50000
o
o (But: 1 ~O)
1 (1995)
10 (1995) 20 (1995) 2b (1995)
Reou1atory Scenario
2c(I995)
3( 1995)
-------
Figure IV-10. Total Uncontrolled and Controlled Toxic VOC Emissions by
Scenar10.
Geographic Area: Los Angeles
1600000 15<4&100 15<48100 1$41&100 15<16100 1s.1~loo 1~61oo
1400000 136 3400
0'\ 1200000
0'\
1 000000
Emissions 800000
( tons/yr )
623500
600000
<100000
200000
0
0 (Be..: 1 QaO) 1( 1995) 10(995) 2e (1995) 2b(995) 2c( 1995) 3( 199$)
Reouletory Scenario
\
-------
Figure IV-11. Total Uncontro11ed and Control1ed Toxic VOC Emissions by
. Scenario.
Geographic Area: Baton Rouge
<400000
0\
-.J
311300
311300
311300
311300
311300
.311300
300000
Emissions 200000
( tons/yr )
201~0
1 QO 300
100000 .
o
o (But: 1 ~O)
1(1995)
la(1995) 2a(I99S) 2b(I99S)
. ROCJu1atory Scenario
2c:(1995)
3( 1995)
-------
Figure IV-12. Tota] Uncontro11ed and Contro11ed Toxic VOC Emissions by
Scenario.
Geographic Area: BaHlmore
'400000
0"\
CX>
3101100
31<1100
31'1100
31<1100
31<1100
3101100
300000
273000
Emissions
( tonslyr ) 200000
123100
100000 .
o (Bou: 1~0)
1 (1995)
10(1995) 20(1995) 2b(I99S)
Reou1atory Scenario
2c(IQ9S)
3( 1995)
o
-------
Figure IV-13. Tota1 Uncontro11ed and Control1ed Toxic VOC Emissions by
Scenario.
Geograph I c Area: Phl1 ade1phf a
100000
145500 115500 115500 145500 145500 145500
140000 130100
0\ 120000
'"
1 00000
Emlssfons ~oooo
( tons/yr )
60000 50100
~oooo
20000
0
0(60u: I~O) 1 (1995) 10(1995) 20(1995) 2b( 1995) 2c (1995) 3( 1095)
Regu1atory Scenarfo
-------
Figure IV-14.. Tota1 Uncontro11ed and Contro11ed Toxic VOC Emissions by
Scenario.
Geographic Area: AI' Areas
3000000
~26oo 2 5-I2() 00 25-12600 2 5-I2() 00 25-12000 25-12600
2500000
-..:I 2117S00
a
200C()00
Emissions 1 500000
(tons/yr)
1 000000
500000
0
0 (eUt; 19aO) 1( 1995) 10(1995) 20 (1995) 2b(1995) 2c(199S) 3 ( 1995)
Regulatory Scenario
-------
:-:-
lu.., I,
o
1111 I,.",
""ru,
I,U
YtMeh.
,r)
1 10\,:
II''',.",
hvrell
It..
y.~lelll
-..J
/-I
IoUI
"
IU",..ry
hv'ell
....
y.Mc '"
h\.1
h
1111 h..ry
St"rc"
I...
y.Mc hI
IoUI
n
1111 "..ry
hvrell
I...
V.~lcIII
'n.1
Ie
II II t '."y
St"'c,,
ItU
V.~lchl
'n.1
I
lu\ h..ry
hv'ul
It..
V.~1c '"
h\.1
I. I.
( I/y,)
1,080
',nl
1,010
',)\1
I,'"
',)\1
),11'
',)\1
I,ur
',111
I,m
. .111
(011
(N'II,,)
"
'f\o,"h
. 01
.01
.11
71.1
.11
11.1
.11
71.1
.11
71.1
..
71. I
I, I
11.1
Table
IV.5
Summary of PM Emissions Reduction (E,R.)
and Associated Costs of Control by Scenario
. COIl
"..
Ultll..
(N'II,,)
101,100
,,'"
m,m
. ,
611,'00
~.. Ana,'..
I. I.
. (1/,,)
11,110
11,110
m.'
m
"'.1
m
ru
m
COl\
(N'I/,,)
. UII
"0.
bll.II..
IM'II,,)
14.1
IU,OOO
1.1.
(II,,)
111,400
III
III
ns
'IS
711
'I'. I
III
"'.4
711
,.,
III
Itlo. lovo.
(01\
(M'lIy,)
101. I
"'
118.'
. COli
"0'
hit 11o.
(M'II,,)
UI
58,110
1\,100
'1\ ,100
11.110
I\.tlO
11,110.. ".1
7C , 100
no.'
71.1
II. I
. I
111,111
II. I
II. I
. I
UI,UI
II. I
11.1
11,110
. I
111,/11
II. I
II. I
11.110
..
111,111
77.1
11.1
11,110
1.1
111.111
11.1
II. I
11,110
101. I
101. I
1Jt. .
m
m
/11.1
m..
107. I
II'
m
III. I
m.'
108.'
ur
m
110. I
108. ,
m
140. I
108.1
m
110.1
101.1
m
111.1
111,100
IU,'IOO
118 .100
111.100
801,100
II.'
II.'
UI.I
III. ,
IJt
II.'
II.'
111..
148. ,
IJt
II.'
II.'
UI.'
III
II.'
II.'
UI.'
1Jt. ,
II.'
II.'
U'
m.'
II.'
II.'
m.I
1.1.
(1/,,)
(III
(M'I/,,)
11,100
/,111
, .11'
/,111
" ,,~
',III
/,111
111\1....
6 COIl
"..
bu. 11o.
(M'II,,)
11.1
II. I
11.1
II. I
II. I
Jr.I
II.' \
II. I
11.1
11.1
Jr.1
'0.1
II. I
'~lltd.lp~l.
. CII\
"..
1.1. (011 bu.II..
(lIy') IM'II,,) (NIII,'1
1.1
II. I
1'.1
I. ,
41.'00 .
1/.1
1'.1
I. .
41./00
Jr. I
11.4
1.1
'1.'00
Jr. I
II. .
I. ,
'1,"0
Jr.1
II. .
'./
11,/10
11.1
H.'
11,100
400
H,OOO
41,100
1,100
1,100
1.100
1,t04
1,100
1,100
I.'
I. I
1.1
10./
II.'
10./
11.8
10.'
II. I
10. ,
10. ,
H./
11.1
10. I
10. ,
II. I
II. .
10./
10./
II. I
111,000
',SlO
"1,1/0
1.1
10.1
n.1
10.1
H.I
10.1
H.I
'.1
'.1
"'I Inll
I.'.
(II,,)
(,,,
(M'II,,)
.
'71.'
110..
114
. enl
I...
"ull..
(M'''y')
m.'
m..
.
III. .
$II
178.1
"'./
II'"
"'.1
"1.1
."
-------
'rable
IV.6
Summary o~ vac Emissions Reductions (E.R.)
and Associated Costs of Control by Scenario
Pho.nll Los Ano.lu 8t ton loua. 8.11 I.or. Phlltd,lphh A II tr..,
4 con 4 cost 4 coli 4 COli 4 con . con
Ir08 '.08 rro. '.08 '.08 '.08
(.1. . Con bu.lIn. 1.1. Cost b'18l1n, 1.1. Cost bu.lln, 1.1. Con b...1 In, I. I. Coli bU.lln. 1.1. Cost b.ft IIn.
Se'.HIo (lIy.) (M'S/y.) (M'S/y.) (lIy.) (M'S/y.) (M'S/y.) (fly.) (M'S/y.) (M'S/y.) (lIy.) (M'S/y.) (M'S/y.) (lIy.) (M'S/y.) (M'S/y.) (1/y.) (M'S/y.) (II'S/y.)
o
Stlllo.HY
Sou.eu 96 503.700 390.7 ZJ,500 1.1 100,000 76.8 36.640 85.9 613,400 U"
lo.d
y.hlelu 10,811 13.8 736,800 731.7 11,"0 11.8 49,900 50.8 40,300 3U '11,300 '1$,4
Tot.1 10,961 13,8 140,000 6Z1.' 36,950 14.9 "',900 Ill.6 16,"0 IlU 1014 ,100 961.4
I
Stllfo...y
Sou.eu 16 ,700 10.0 10.0 780,600 650 759.8 86,000 10.4 9.3 I1Z ,000 1JJ.9 51. I 48,900 98.7 12.3
lo.d
y,hlelu 141,700 1IZ.7 98.6 411,800 483.8 746.5 74,440 76.1 17.3 7.,400 92.' 41.6 66 ,000 16.9 31,1 .
ToUI 108.6 506.3 11,6 98.1 49.4 17S4 784.6
It
Stu lon..y
Sou.eu 11,110 1l,5 1l.5 185,500 6S1 760.8 90,700 1$.1 14. 1:1,'00 136.5 59.1 50,150 99.1 13.7
-.:)
I\) Ao.d
y,hlelu 141. 700 117.7 98.6 411,800 '8). 8 746.5 7. ,440 76.1 IU 9,'00 97.' '1.' 66 ,000 16.9 31.1
Tot.1 111.1 501.3 7U 101.J 50.3 1165,7 796,1
Zt
Stlllon,.y
Sou.eu 11,440 11.7 11.7 19Z ,000 661.6 711.. m,900 n.1 70 118,100 136.5 59.1 50,800 99.3 13.4
Ao.d
Y.nlelu 141,700 117.7 98.6 '11,800 483.8 746.5 74,440 76.1 lib I, .400 92.4 41.6 66 ,000 16.' 37.1
ToUI 111.8 SZJ.9 31 o. 101. 3 50.5 1789.1 810.3
7b
SUt 10."y
Sou.eu 71,440 13.7 ~3.1 197.000 661.6 711.4 1S5, 900 71.1 10 118,200 136.5 59.1 50 ,800 99.3 11.4
Ao.d
Y,hle Its 147,700 114,3 100.5 411,800 483.8 746. S H,I45 78.' 14.6 19 ,250 91.5 41.1 66,100 11.1 31.9
Toul lU.l SZJ.9 34.6 101.4 SI.3 1795,3 US. t
Ie
St.t 10.Hy
Sou.eu 16, no ll.' 1l.4 8JO,800 1$9., 369.1 706,500 78.1 71 183,100 141.5 64.1 53,600 104.' 11.5
Ao.d
y.nlelts 147,100 114.3 100.5 411,800 '81.8 746.5 25,145 lB.' 14.6 19 ,750 93.5 Il.l 66 ,100 71.7 37.9
Toul 1l7.9. 615.1 91.6 101.. U.4 1961.' m
3
St. II onHy
Sou.eu 31,810 47.3 '1.3 883,600 1097 101.8 758,600 46.6 '5.5 01,500 ISt.4 87.6 56,400 lIS 79. I
AO'd
y.n I e Its 11l,100' 11.. J 100.5 41) ,800 483.8 146.5 75,14 5 78.4 14.6 ",750 9J.5 47.1 66,100 71.1 37.9
Iot,1 147.8 948.1 60.1 m.3 61.0 7m,t UU.S
-------
~atle IV.7
P!1 2nd VOC Emission Control Cost Increases
, ( . . 2 ~. )
in all ::'i ve Hreas ''', .j:/yr
PH VOC Both pollutants
Compared Compared Compared Compared Compared Compared Compared Compared Compared
to to to to to to to to to
Scenario baseline Scenario 1 Scenario la baseline Scenario 1 Scenario la baseline Scenario 1 Scenario 1a
565 784.6 1349.6
\a 644.7 79.7 796.3 11. 7 1441 91.4
2a 645.2 80.2 .5 820.3 35.7 24 1465.5 115,9 24.5
2b 645.2 80.2 .5 825.9 41.3 29.6 1471. 1 121. 5 30.1
2c 647.0 82.0 2.3 993 208.4 196.7 1640 290,4 199
-..J 3 650.6 85.6 5.9 1343.5 558.9 547.2 1994,1 644.5 553.1
w
-------
associated with each scenario relative to the following:
. baseline control costs (1980),
. Scenario 1 (1995), and
. Scenario 1a (1995).
As in previous tables, the cost data are expressed in
millions of 1984 dollars per year and include both stationary and
mobile sources.
. Appendices e and D (Radian) show the RIM outputs for
toxic PM and voe emissions for stationary sources. Appendix e
(Radian) presents results aggregated into the 14 major source
categories studied as defined in Table IV.8.
Two tables are presented for each scenario for each area
for toxic PM and voe emissions individually. . The first is the
emission projections output of the RIM emission projection module
and the second presents the emission control costs associated
with the degree of control required of each source type.
Appendix D (Radian) presents the same information as .
Appendix e (Radian) except that the major source categories are
disaggregated to the see level. To identify the name of the see,
Appendices A and B (Radian) present cross ref~rence definitions.
As previously indicated, projections for road vehicles were
performed using the MOBILE3 projection model and, thus, do not
appear in the RIM outputs.
74
-------
T a bl e IV. 8
Major Source Categories Studied
..c;..9.t~.g.Q.r.L.N.9....
1
2
3
4
5
6
7
8
9
10
1 1
12
13 (not presented)
14
.M..9j.Q.r_So'yJ:.Q~..c;..s.t~.g.Q.r.Y
Steril iz ers
Dry cleaning
Util i ty boil ers
Waste oil combustion
Iron and steel manufacturing
Refractory manufacturing
Non-ferrous and secondary metals
pe tr 01 eum ref ini ng
Gasoline marketing
Chemical manufacturing
Sol vent use
Wood smoke
Road v ehicl es
Other source types
75
-------
V EXPOSURE MODELING
A. MODELING METHODS
Assessing the cancer risks posed by a suspected airborne
carcinogen requires three basic pieces of information: 1) an
estimate of the carcinogenic potency of the pollutant in
question, 2) an estimate of the ambient concentration to which an
individual or a population is exposed, and 3) an estimate of the
exposed population. Described below are the methods for
developing and combining the above three pieces of information in
order to estimate cancer incidence in the five study areas.
The carcinogenic potency of a compound represents the.
probability of contracting cancer from a lifetime (70 years)
exposure to a given concentration of that pollutant. The
carcinogenic potency estimates used in this study were developed
by EPA's Carcinogen Assessment Group. They are listed in Table
V. 1 .
Estimates of ambient concentrations to which people are
exposed were estimated using the Systems Applications Human.
Exposure and Risk (SHEAR) model; a set of computer codes and data
files designed to estimate patterns of pollutant concentrations
and of related measures of health risk due to all identified
sources of potentially hazardous species in a designated modeling.
region. Another model, the Office of Toxic Substances fate and
transport model known as GAMS, w~s considered as an option for.
performing the dispersion modeling portion of the study. It
performs basically the same operations as SHEAR. Both models
produce annual average concentration estimates, use EPA's
Stability Array (STAR) data to represent the climatology of each
city, and estimate exposure using 1980 block group/enumeration
district populations. (Neither SHEAR nor GAMS is an EPA approved
model.) We chose SHEAR for this study because it appeared to do
a better job of modeling area source emissions and of identifying
sou~ce culpability.
In SHEAR, two basic dispersion algorithms are used: a
plume algorithm for computing concentration patterns resulting
from a single source (point sources) and a box algorithm for
computing concentration patterns resuiting from an urban-wide
distribution of a single source type (area sources). The
dispersion zone for each modeled point source is limited to a
circle of a 20 kilometer radius about the source location. This
limitation should have minimal effect on study results because
cOhcentrations from most sources will be small at 20 kilometers. .
Exposure to air toxics in SHEAR is computed for
residential population patterns for annual-average
concentrations. No account is taken of varying source strengths
76
-------
Ta bl e V. 1
Unit Risk Values Used in the Air Toxics Controllability Study
.£.Q.l.l.Y.t.s.D.t
Benz ene
JJ.Dj.t_.Bj~.k.!
4.3 x 10-3
8.2 x 10-6
Arsenic
Chloroform
5.0 x 10-1**
1 . 5 x 1 0 -5
1 .0 x 10-5
1 .2 x 1 0 -2
7.0 x 10-6
6.1 x 10-6
8.9 x 10-7
3 .3 x 10-4
1.7 x 10-6
4 . 1 x 10-6
2.6 x 10-6
3.6 x 10-4
Products of Incomplete Combustion
Carbon Tetrachloride
Chromium
Ethylene Dichloride
Formal dehy de
Gasol ine Vapors
Nickel
Perchloroethylene
Trichloroethylene
Vinyl Chloride
Ethy lene Oxi de
. ;
*The unit risk value is the estimated probability of
contracting cancer as the result of a constant exposure over 70
years to an ambient concentration of one microgram per cubic
meter.
**. .
The listed unit risk for PIC was applied to all source
~ategories except5road vehicles. For road vehicles, a unit
risk of 3.1 x 10- was applied to exhaust particulate to
estimate incidence from this category for PIC.
77
-------
or patterns, or varying population patterns within a year. In
this study, atmospheric transformation of toxic compounds has
been ignored.
In the point source model, the concentrations from each
source are computed on a separate polar grid centered on that
source. Population is defined on an irregular grid of block
group/enumeration district centroids. To define all values at a
common set of points with a minimum of interpolation, SHEAR adds
interpolated values of concentration from each polar grid to the
grid of centroids.
SHEAR was applied in the five study areas to estimate
expected increases in cancer incidence resulting from air toxics
exposure in the base year (1980) and for six scenarios for 1995.
All point sources for which Pechan had coordinates and stack
parameters (from NEDS) were modeled in SHEAR as point sources.
Area sources and point sources for which Pechan had no stack or
locational information were modeled as area sources. Area source
emissions were scaled by the population density in each block
group/enumeration district.
Because the SHEAR was not designed to be run for large
urban areas with many point sources, a large number of
pollutants, and many scenarios, some changes were made to the
model to make it more efficient for the purposes of the present
analysis. Rather than running the SHEAR separately for each
scenario and pollutant, we ran it once assuming 100 tons per year
of pollutant emissions from each major point source. For each .
point, the output of SHEAR was saved in an intermediate file for
later use. This intermediate file contained the cumulative'
population exposure (microgram-persons/year) for each modeled
point. These values were then used to estimate population
exposure to individual pollutants by multiplying them by the
ratio of actual emissions to modeled emissions (100 TPY). With
14 pollutants and six scenarios, these methods saved considerable
computer resources.
Because the cost and execution time of SHEAR is linearly
related to the number of sources modeled, it was necessary to
establish a minimum emission limit for each pollutant in each
study area. ~hese limits were set to minimize the number of
sources to be modeled while maximizing the percentage of covered
emissions within each study area. For some pOllutants, there
were so few emitters that all sourc~s within a study area could
be modeled (e.g., carbon tetrachloride). On the other hand, some
pollutants had so many sources that a minimum emission limit was
needed to avoid having hundreds of sources included in the
modeling of that pollutant (e.g., formaldehyde).
Table V.2 shows the minimum emission limits used for each
pollutant by study area. Point sources with emissions below the
78
-------
Table V.2
Minimll11 Fmission Limits Used for
Identifying Major Point Sources (TPY)
Los Baton
Pollutant Phoenix Angeles Rouge Ba~timore Philadelphia
Arsenic 0.0 0.2 0.0 0.05 0.1
Benzene 0.0 1.0 0.0 0.0 0.0
Carbon Tetrachloride 0.0 0.0 0.0 0.0 0.0
Chloroform 0.0 0.4 0.0 0.0 0.0
Chraniun 0.0 0.1 0.0 0.05 0.05
Ethy lene Di chl or i de 0.0 0.04 0.0 0.0 0.0
Ethylene Oxide 0.0 . 0.0 0.0 0.0 0.2
Formaldehyde 1.0 70.0 100 20.0 1.0
Gasoline Vapors 0.0 40.0 0.0 100 0.0
Lead 0.0 10.0 0.0 1.0 1.0
Nickel 0.0 0.1 0.01 0.1 0.1
Perchloroethylene 0.0 50.0 0.0 0.0 1.0
PIC 0.0 0.0 0.0 0.0 0.0
Trichloroethylene 0.0 1.0 0.0 0.0 1.0
Vinyl Chloride 0.0 0.0 0.0 0.0 0.0
79
-------
minimum emission limit were modeled as area sources.
Because accuracy in the stack data used in the dispersion
modeling is as critical as the emission rates in estimating
ground-level concentrations, considerable effort was taken in
performing quality control checks on the point source stack
parameters. Steps taken to ensure reasonable values for stack
parameters were as follows:
1. UTM coordinates that were widely different for the same
plant were corrected.
2. Plants in Riverside and San Bernadino Counties which were
outside the South Coast Air Basin were deleted.
3. The stack heights and diameters were reviewed for
compatibility.
4. Sources with stacks taller than 100 feet should be fuel
combustion sources and have a stack gas temperature of
more than 200 degrees F.
5. If temperature, height, and diameter are zero or missing,
values are assigned.
6. Default values for exhaust gas flow rate and stack
temperature for utility and industrial boilers are based
on the type of fuel burned.
7. In cases where plume height was estimated but needed
stack data were missing, plume height was used to
estimate stack height. Other stack parameters were then
estimated using plume rise formulas. .
B. RESULTS
Modeling results shown in Table V.3 indicate the
estimated annual incidence summed over the five study areas will
drop from 802 in the base year to 584 in 1995 under controls
expected to be applied as a result of the current regulations to
control criteria pollutants and current NESHAPs. Thus, it is
estimated that incidence will be reduced by 27 percent over this
period. This 27 percent reduction in incidence is expected to
result largely from a 40 percent reduction in road vehicle
emitted toxics (primarily PIC) and a 13 percent reduction in PIC
from residential wood combustion.
Scenario 1a in Table V.3 adds to the current regulatory
program the anticipated effects of possible new NESHAP
initiatives. Because the effects of reduction in chromium
emissions from new NESHAPs are significantly higher than those of
80
-------
TableV.3
Summary of Estimated Annual Cancer Incidence
by Scenario*
Estimated Percentage Addi ti onal
Annual Reduction Percentage
Incidence from Existing Reduction
Existing Conditions (1980) 802
1QQS Scenarios
1. Criteria Pollutant Program
Plus Existing NESHAPs 584 27.2% 27.2%
1a. With New NESHAP Initiatives
A. For Chranium Only 557 30.6% 4.6%
B. All Pollutants. 531 33 ~7% 4.5%
2a. With most stringent controls
on new sources 529 .34.0% 0.6%
2b. With most stringent control s
on road vehicles 528 34.1% '0
2c. With most stringent controls
on replacement sources 497 38.0% 5.9%
3. With most stringent controls
on all (new, repl acement
and existing) sources 455 43.2% 8.4%
* Incidence numbers presented above assume constant exposure to pollutant
levels during a 70 year period. Cancer unit risk values applied were
obtained fran EPA's Carcinogen Assessment Group.
81
-------
any other pollutant, the chromium results are listed separately
in Table V.3. The 9 percent incidence reduction from new NESHAP
ini tiatives comes largely from reducing chromium emissions from
chrome plating operations and reducing ethylene oxide emissions
from sterilizers. The effects of other new NESHAP controls are
minimal.
Scenarios 2~ and 2b in Table V.3 show that addition of
more stringent controls on new sources than what is currently
expected brings little additional reduction in incidence; only
more stringent controls on replacement and existing stationary
sources (Scenarios 2c and 3) reduce incidence significantly
beyond what is expected through current regulations plus NESHAPs.
Figure V.1 is a graphic presentation of the information
in Table V.3. It illustrates how reductions in cancer incidence
will change as more stringent control programs are implemented.
City-to-city variations in incidence are large, as
illustrated by Figure V.2. The expected incidence in Los Angeles
is greater than that of the other four study areas combined. If
incidence is computed on a per capita basis, the rankings change
somewhat. ~hiladelphia and Baltimore have the highest per capita
incidence, followed by Los Angeles, Phoenix, and Baton Rouge.
These per capita incidences are illustrated in Figure V.3.
. .
To isolate the major sources and pollutants coritributing
to incidence levels, the top five source category-pollutant
combinations for each city were examined. This information is
presented in Table V.4. Products of incomplete combustion
emitted by road vehicles category is the top contributor to
incidence in all five study areas. This contribution varies from
46 to 79 percent depending on the study area. (Table V.4 values
are expressed as percentages in Table V.5.) Products of
incomplete combustion from. wood burned in fireplaces and wood
stoves also contribute significantly to total risk in all five
study areas. Wood smoke-PIC contributions vary from 6 percent in
Los Angeles to 36 percent in Baton Rouge.
In total, PIC emitted by road vehicles and residential
wood stoves and fireplaces represents 81 percent of all incidence
in the five study areas. An additional 10 percent of total
incidence results from chromium emissions. Significant chromium
emitters include chrome plating operations, cooling towers, and
refractory manufacturing. 'Ethylene. oxide, benzene and
formaldehyde are the only other pollutants which contribute more
than 1 percent to the total incidence. Their contributions are
2.6,2.4.and 1.1 percent, respectively.
Because PIC and chromium so dominate the incidence
estimates, it is useful to also examine the results with these
pollutants removed. A summary of estimated annual incidence by
82
-------
Figure V.I
Total Annual
Incidence by Scenario
INCIDENCE
1000
900
802
800
700
Q)
w
600
500 497
400
300
200
100
o
~9 0 1"\9 L1 t--?S ee.9 \ e.9 ee.S
0\ \i \ 0 ~\J 1'9':) ~~SP sour 'J e.\"\ \ e Sour
\'\~ co~ Curre "" ~e.'d sc./~e.~ "" v.,Sc./ (~,e.? \. £.j..\s\J.
'C.)l.~9\J\ ~ "" ~ ~Sc. "" y.,sc./
MSC = most stringent controls
. ,ree.S
Sov
-------
ANNUAL INCIDENCE
1000
900
800
700
co
.l:
600
500
400
300
200
100
o
Phoenix
Ficure V.2
City-Specific Analysis
Existing Conditions
494
11
Los Angeles Baton Rouge
132
Baltimore Philadelphia
-------
Fit;ure V.3
City-Specific Analysis
Existing Conditions
ANNUAL INCIDENCE PER CAPITA (10-6)
100
90
80 77 78
70
CD
V1
60
50
40
30
20
10
0
Photm i x
Los Angeles Baton Rouge
Baltimo~e Philadelphia
-------
Table V.4
Annual Incidence Under Existing Conditions'
Top Five Source Category-pollutant Combinations
Phoenix Los Angel es Baton Rouge Bal timore Philadelphia
,
Road Vehicle-PIC 43 Road Vehicle-PIC 388 Road Vehicle-PIC 6 Road V ehicl e-PIC 46 Road Vehicle-PIC 65
Wood Smoke-PIC 17 Wood Smoke-PIC 32 Wood Smoke-PIC 4 Wood Smoke-PIC 32 Chrome Plating-CR 28
All Other-CR 3 Chrome Plating-CR 28 All Other-CR 1 All Other-CR 8 Wood Smoke-PIC 15
Road Vehicle-Benzene 1 Sterilizers-EtO 18 All Other-Carbon Tet. <1 Regulatory Mfg.-CR 6 Residential Coal-PIC 15
Gasoline Marketing- 1 Road Vehicle-Benzene 13 Road V ehicl es-Benz ene <1 Chern. Mfg.-CR 1 Road Vehicle-Benzene 2
Gas Vapors
Other Other 15 Other 0 Other 6 Other 7
ex> Total 66 Total 494 Total 11 Total 99 Total 132
0'\
'Incidence numbers pr~sented above assume constant exposure to pollutant levels during a 70-year period.
were obtained from EPA I S Carcinogen Assessnent Group.
Cancer unit risk values applied
-------
Table V.5
Incidence Under Existing Conditions
Top Five Source Category-pollutant Combinations
Percentage by Study Area
Source Category - Pollutant Combinations Phoenix Los Angeles Baton Rouge Baltimore Philadp.lphia
Road Vehicle-PIC 65 79 55 46 49
Wood Smoke-PIC 26 6 36 32 11
Chrane Plating-CR 6 21
(» Refractory Mfg.-CR 6
-...:J
Steril izers-Eto 4
Road 'Vehicle-Benzene 1 3 2
All Other-CR 5 9 8
Gasoline Marketing-Gas Vapors 1
Chemical Mfg. -CR 1
Residential Coal-PIC 10
% Top Five Are of Total 98 98 100 93 93
-------
scenario, with PIC and chromium not included, is shown in Table
V.6. Results show somewhat higher reductions in these remaining
pollutants for most scenarios than those shown in Table V.3.
While the reduction in estimated annual incidence as a result of
the criteria pollutant program is less than shown in Table V.3,
the new NESHAP initiatives bring a significantly greater
incidence reduction with PIC and chromium excluded. Other more
stringent control programs provide slightly lower reductions if
PIC and chromium are excluded from the analysis, than if they are
included, but this may be a product of the higher reductions
already achieved under less stringent programs.
Table V.7 shows the estimated annual cancer incidence by
pollutant summed for all five study areas. Results for existing
conditions (1980) plus all six 1995 control scenarios are shown.
Because scenario 1 is considered the most likely 1995 case, it is
instructive to compare it to the existing case. Comparing the
existing case with Scenario 1 for 1995 shows that reductions in
PIC are largely responsible for the decrease in estimated cancer
incidence. In fact, the reduction in incidence from PIC is 35
percent, while the estimated overall reduction in incidence from
the existing case to scenario 1 is 27 percent. A 34 percent
increase in estimated incidence from exposure to chromium offsets
some of the PIC reduction. No other pollutants show incidence
increases of more than 15 percent. Besides PIC, there are three
other pollutants that are. expected to be relatively well.
controlled (more than 20 percent decline) in the 1980-1995
period. These pollutants are benzene, trichloroethylene, and
gasoline vapors. However, these reductions are largely
overshadowed by the PIC and chromium changes.
Further controls, in the form of new NESHAPs (scenario
1a), decrease the chromium incidence figure to 1 percent below
the baseline. New NESHAPs also have the potential to
significantly reduce expected incidence from ethylene oxide,
trichloroethylene and gasoline vapor emissions. Incidence
associated with ethylene oxide exposure is estimated to be
reduced by 96 percent as a result of new NESHAPs. . This
represents 42 percent of the incidence change between scenarios 1
and 1 a.
What little change there is between scenarios 1a and 2a
(the addition of more stringent controls on new sources) is in
the incidence estimate for PIC. Adding more stringent controls
on road vehicles (scenario 2b) prod~ces no perceptible change in
incidence; estimated incidence from benzene and formaldehyde is
~educed only slightly.La~ger reductions in incidence are
expected from controls on replacement sources (scenario 2c).
Almost all of the compounds under study are reduced somewhat by
this control strategy. A 7 percent decrease is estimated in
incidence from PIC, when compared with scenario 2b. Decreases in
incidence from other pollutants are much smaller.
88
-------
Table V.6
Summary of Estimated Annual Cancer Incidence
by Scenario*
All Pollutants Except PIC and Chraniun
Estimated Percentage Addi ti onal
Annual Reducti on Percentage
Incidence from Existing Reduction
Existing Conditions (1980) 61.3
1 qql) Scenarios
1. Criteria Pollutant Program
PI us Exi sting N ESHAPs 50.3 18% 18%
1a. With New NESHAP Initiatives 25.3 59% 50%
2a. With most stringent co ntrol s
on new sources 25.1 59% 0.5%
2b. With most stringent controls
on road vehicles 25.0 59% 0.5%
2c. With most stringent controls
on replacement- sources 24.4 60% 2.4%
3. With most stringent controls
on all (new, repl acement
and existing) sources 23.4 62% 4.2%
* Incidence numbers presented above assume constant exposure to pollutant
levels during a 70 year period. Cancer unit risk values applied were
obtained fran EPA's Carcinogen Assesffilent Group (CAG).
89
-------
Table V.7
Expected Annual Cancer Incidence by Pollutant Assuming Increasing Regulation
Scenario Scenario Scenario Scenario Scenario Scenario
Existing( 1) 1 1a 2a 2b . 2c 3
~ ~ ~ ~ ~ ~
Pollutants Incidence(2) Incidence Reduct. en Incidence Reduct. Incidence Reduct. Incidence Reduct. Incidence Reduct. Incidence Reduct.
PIC 663 429 35 429 35 426 36 426 36 397 40 358 46
Chraniun 78 104 (34) 77 1 77 1 77 1. 76 2 74 4
Berrzene 19 9 55 8 57 8 57 8 58 8 58 8 59
Formaldehyde 9 7 19 7 19 7 19 7 20 7 21 7 24
Arsenio 1 1 (7) 1 8 1 9 1 9 1 28 1 58
Ethylene Oxide 20 23 ( 15) 1 96 1 96 1 96 1 96 1 96
Carbon Tetra- 2 2 0 2 0 2 0 2 0 2 0 2 0
chloride
TCE 2 1 31 <1 74 <1 74 <1 74 <1 74 <1 74
Gas Vapors 3 2 24 1 57 1 55 1 55 1 57 1 58
PERC 4 3 8 3 15 3 18 3 18 3 21 3 26
Nickel <1 <1 1 <1 14 <1 15 <1 15 <1 36 <1 79
\.D £DC <1 <1 0 <1 1 <1 1 <1 1 <1 1 . <1 1
o
Chloroform <1 <1 0 <1 0 <1 0 <1 0 <1 0 <1 0
Vinyl Chloride ~ ~ -.0. ~ -.0. ~ -.0. ~ -.0. ~ -.0. ~ -.0.
Total 802 584 531 529 528 497 455
(1) Existing: Existing Regulations
1 : 1995 Current Regulations
1a : 1995 Current Regulations + NESHAFS
2a : Scenario 1a + most stringent controls on new sources
2b : Scenario 2a + most stringent controls on road vehicles
20 : Scenario 2b + most stringent controls on replacement sources
3 : ~st str1ngent controls on all sources
(2) Inc1dence 15 the total nunber of excess cancers that may occur each year.
(3) Percentage reduotion fran existing conditions.
-------
Further significant reductions in PIC-related incidence
are observed in scenario 3. In fact, 93 percent of the incidence
change from scenario 2c to scenario 3 is a result of reductions
in PIC emissions.
Perhaps the most useful insights that can be gained from
Table V.7 are that there are a number of toxics for which there
seems to be little expectation of reduction in incidence, even
with the most stringent controls. Chromium, carbon
tetrachloride, ethylene dichloride, chloroform, and vinyl
chloride show little change from 1980 conditions. Because
chromium sources contribute as much as 16 percent to the total
incidence in the five study areas (scenario 3), the fact that it
cannot be reduced beyond the baseline level is a more serious
problem than for the other pollutants. This may imply that
substitution of a less toxic material or shutdown of operations
with significant chromium emissions are the only effective
control techniques.
It is also useful to examine how different source.
categories are affected by the various control scenarios. Table
V.8 presents this information for the 14 source types that
contribute the most to incidence in the base year. Table V.8
shows that road vehicle emissions are expected to be reduced
significantly (41 percent) under the criteria pollutant program
but that little further reduction is feasible with current
technology. Conversely, residential wood combustion related
incidence is not reduced substantially unless existing and
replacement stoves and fireplaces are controlled. Thissuggests
that the market for wood stoves and fireplace inserts is
saturated, or nearly so.
For chrome plating facilities, Table V.8 shows that new
NESHAPs (scenario 1a) are needed to keep estimated incidence from
this category from increasing substantially above what it is in
the base year. Another important chromium emitter,refractory
manufacturing, is not expected to be controlled enough to offset
the effects of growth, even under the most stringent control
scenario. This suggests that most of the chromium sources in the
refractory manufacturing plants in the study areas are already
well controlled.
Ethylene oxide emissions from sterilizers are estimated
to be controlled by 99 percent via a new NESHAPs initiative
(scenario 1a). Thus, sterilizers go from contributing to 4
percent of total incidence in scena~io 1 to being a
non-significant contributor to total incidence in scenario 1a.
C. SENSITIVITY ANALYSIS
As mentioned in the introduction to this report, for
91
-------
Ta bl e V. 8
Expected Annual Cancer Incidence by Pollutant Assuming Inc~easing Regulation
Exi sti ng (1) 1a 2a 2b 2c 3
$ $ $ $ , $ $
Pollutants IncidenceC21 Incidence Reduct. en Incidence Reduct. Incidence ~ Incidence Reduct. Incidence Reduct. Incidence ~
Road V ehicl es 571 338 41 338 41 338 41 338 41 338 41 338 41
Wood Smoke 100 87 13 87 13 84 16 84 16 56 44 17 83
Chrane Plating 69 94 07> 67 2 67 2 67 2 66 4 65 6
Sterlliz er s 20 23 (16) <1 99 <1 99 <1 99 <1 99 <1 99
Residential Coal 16 16 ( 1) 16 (1) 16 (1) 16 (1) 15 4 14 9
Refractory Mfg. 6 7 (20) 7 (20) 7 (20) 7 (20) 7 (20) 7 (20)
Chanical Mfg. 2 2 3 2 13 2 14 2 14 2 14 2 15
Iron & Steel 1 1 (17) 1 (7) 1 (6) 1 (6) 1 (2) <1 8
Oil Combustion 1 1 4 1 40 1 40 1 40 1 40 1 50
Utility Boilers 1 1 7 1 7 1 9 1 9 1 16 <1 77
Sol vent Use 3 3 25 2 55 2 55 2 55 2 55 2 55
Petroleun Refining 1 1 10 1 10 <1 17 <1 17 <1 33 <1 85
Gasoline Marketing 3 2 26 1 57 1 55 1 55 1 57 1 59
\0 Dry Cl eani ng 2 2 (1) 2 ( 1) 2 5 2 5 2 10 2 18
I\) All Others -6 J -5 -:I. -:I. -5 J
Total 802 584 531 529 528 497 455
(1) Existing = Existing Regulations
1 = 1995 Current Regulations
'1 a = 1995 Current Regulations + NESHAfS
2a = Scenario 2a + IIX>st stringent controls on new sources
2b = Scenario 2a + most stringent control s on road vehicles
20 = Scenario 2b + most stringent controls on replacement sourQes
3 = I-bst stringent controls on all sources .
(2) Incidence is the total number of excess cancers that may occur each year.
0) Percentage reduction fran existing conditions.
-------
chromium, the major uncertainty results from the difficulty in
determining whether emissions are in the trivalent or hexavalent
form. Only the hexavalent form has been proven carcinogenic;
there is insufficient evidence to determine if the trivalent form
is also carcinogenic. For the purposes of this analysis, total
chromium emissions were assumed to be as carcinogenic as the
hexavalent form. Recent evidence from source tests shows that
the percentage of chromium that is hexavalent varies by source
type and that this percentage may be as low as 1 percent in
refractories. Thus, a sensitivity analysis was performed to
examine how the analysis results might change according to
assumptions about the percentage of chromium emissions that are
hexavalent. In this sensitivity analysis it has also been
assumed that trivalent chromium is not carcinogenic. (Note that
this assumption differs from what was assumed for the body of the
analysis in this report.)
Table V.9 summarizes the results of the sensitivity
analysis across all five study areas. If only 10 percent of
chromium emissions are carcinogenic, the total annual estimated
incidence drops by 9 percent, or 71 excess cancer deaths. A 1
percent hexavalent assumption reduces the estima-ted incidence by
another percentage point. Because chromium is not expected to be
controlled below what emissions are in the baseline, the current
regulatory program and new NESHAPs become more effective in
redu~ing overall incidence with chromium.making less of a
contribution. The effectiveness of scenario 1 goes from a27
percent reduction if all chromium is carcinogenic to a 34 percent
reduction if 1 percent of chromium is carcinogenic.
Because the contribution of chromium to annual incidence
differs by study area, the effect of assuming that only 10
percent of chromium emissions are carcinogenic effects the
results in each city differently. For existing conditions,
excess cancer deaths are reduced by 20 percent in Philadelphia
and 15 percent in Baltimore if only 10 percent of chromium
emissions are carcinogenic. For the other three study areas, .
incidence estimates are only lowered by 5 percent with the change
in assumptions.
It should be noted that the sensitivity analysis
presented here only attempts to examine incidence changes in a
crude fashion. Recent source tests indicate that the percentage
of hexavalent chromium emissions differs by source type; the
situation was modeled as if all sources emit the same percentage
of hexavalent chromium. To have tried to capture the different
emission patterns from each source type would probably. have been
unreasonable given the preliminary nature of the available
emission measurements. In any case, it is recommended that
future work take into account the most recent information on
specific chromium compounds being emitted.
93
-------
Table V.9
Sensitivity of Estimated Annual Cancer Incidence
to Assumptions About the Percentage
of Chranilll1 Emissions that is Hexavalent
Estimated Annual Incidence
100%
hexavalent
10%
hexavalent
1%
hexavalent
Existing Conditions (1980)
802
731
725
1qqc:; Scenarios
1. Criteria Pollutant Program
Plus Existing NESHAPs
1a. With New NESHAP Initiatives
584
531
490
462
480
455
* Incidence numbers presented above assume constant exposure to pollutant
levels during a 70 year period. Cancer unit risk values applied were
obtained fran EPA's Carcinogen Assessment Group.
94
-------
D. SUMMARY AND CONCLUSIONS
The results presented in the preceding section can be
summarized as follows~
(1) In the cities studied, PIC emitted by road vehicles
and residential wood combustion predominate. The current
regulatory program is expected to reduce road vehicle emitted PIC
by 40 percent by 1995 and this emission change is expected to
reduce excess cancer deaths by 220 per year. New initiatives to
control PIC beyond those included in current regulations are not
expected to be successful in further reducing the expected
incidence from this source type, however.
(2) PIC emitted by residential wood combustion is not
expected to be controlled appreciably by the current regulatory
program; significant reductions from this source type are not
evident unless controls on existing wood stoves and fireplaces
are installed. Previous work in air pollution problems
associated with wood stoves has focused on the Pacific Northwest
and New England. Study results presented here indicate that
major urban centers outside these regions may have residential
wood combustion related problems as well.
(3) Chromium emitted by chrom~ plating shops will lead to
increased incidence unless controls beyond those expected from
the current regulatory program are imposed. Even the most
stringent available controls for this source type, though, will
not reduce incidence below current levels. .
(4) A new NESHAPs to reduce ethylene oxide emissions from
sterilizers has the capability of reducing incidence associated
with this source type to negligible levels.
(5) Area sources represent most of the incidence reported
in this study. This may result from biases associated with the
five cities studied here, but it is also probably associated with
the proximity of area sources to the affected population. In
other words, if major point source emitters are isolated from
people's residences, their emissions are likely to be well
dispersed before they reach residential areas. This implies
that proper land use planning may help reduce the risk of
contracting cancer through airborne exposures to toxic compounds
emitted by stationary sources.
95
-------
VI HOT SPOT ANALYSIS
A. MODELING METHODS
The previous chapter reported on expected cancer
incidence over a 70-year period in each of the five study areas,
but another critical issue is identifying the maximum hazard to
which people might be exposed in any of the five cities. To
investigate that issue, a separate modeling analysis was
performed. The methodology used included modeling all major
point sources in each area using CDM and laying out a receptor
grid network that concentrated receptors near major point sources
(to ensure that points of maximum hazard would not be
overlooked). Receptors were spaced farther apart in places with
no point sources.
To estimate the hazard to the population of exposure to
air toxics,the population data in the SHEAR model (by block
group/enumeration district) were used. Each block group/
enumeration district in the population file has identifying
coordinates (latitude and longitude). These coordinates were
used to associate each block group with the nearest receptor.
Then, the maximum hazard was calculated as the sum of air toxic
concentrations times unit risks. The sum of the hazard at each
receptor was then compared with the number of persons exposed at
each receptor. .
B. RESULTS
The results of this hot spot analysis are presented
separately for each of the five study areas.
1. Phoenix
Figure VI.1 shows risk level as the ordinate (vertical
axis) and population as the abscissa (horizontal axis). . This
graph shows that as much as 10 percent of the people in Phgenix
(approximately 100,000) are exposed to a risk of 1.6 x 10- .
or more, and that one-half of the6Population in Maricopa County
is exposed to a risk of 0.5 x 10- or more. Figure VI.1 also
shows that there is little reduction in point source contributed
risk from current regulatory programs. .
The differences among hazard levels from one population
percentile to another ~n Figure VI.1 is also instructive and
shows that while there is somewhat of a hot spot in Phoenix, the
hazard level decreases gradually with increasing population,
which indicates that the distribution across the city is
reasonably uniform. Other cities have a more L-shaped curve,
which indicates more of a hot-spot phenomenon. It should be
96
-------
Figure VI.l
Key Air Taxies Point Source Hazards
Phoenix
1.7
1.6
1 .5
1,.4
1.3
,--... 1.2
to
( 1.1
0
,.... 1
*
..........
\D (l) 0.9
~ >
(l) 0.8
'U 0.7
I...
0
N 0.6
0
I 0.5
0.4
0.3
0.2
0.1
0
10 20 30 40 50 60 70 80 90 100
I2:Z] Base
Population (percentile)
cs:SJ NAAQS/NESHAPS
-------
noted when reviewing the Phoenix results, though, that Phoeni~
has few point sources and, therefore, would be expected to have
the lowest hazard of the study areas from point source
contributions.
Figure VI.2 is a three dimensional plot of hazard versus
receptor location for Maricopa County. The vertical axis on this
graph, HSUM, represents the sum of the hazards of exposure to all
14 toxic compounds under study (see Table V.1 for a list of these
compounds). HSUM is calculated at each receptor by multiplying
each pollutant's estimated ambient concentration by its unit risk
and summing these values. Figure VI.2 shows that, while there is
some spatial variation in risk within the county, that the
maximum risk is not particularly high. The point of maximum risk
in the western part of Maricopa County is the result of
formaldehyde emissions from an open burning dump. This point of
maximum risk is a considerable distance from downtown Phoenix,
however.
2. Los Angeles
Figure VI.3 shows that 40 percent of th5 population in
Los An gel e s i sex p 0 sed to a h a z a r d 0 f 1. 0 x 1 0 - - 0 r m 0 ref rom
exposure to point source emitted air toxics, and that this hazard
is expected to decrease only slightly by 1995 under current.
regulatory programs. Los Angeles is the only one of the study
areas where incidence is expected to increase between now and
1995. This may imply that regulations in the South Coast Air
Basin are already more stringent than in most areas, and th~t
further reductions via traditional controls are not available.
If Figure VI.3 is compared with the comparable figure for Phoenix
(Figure VI.1) there are two noticeable differences. One
difference is the level of risk; it is more than an order of
magnitude higher in Los Angeles than in Phoenix. The second
difference is the shape of the curves; the Los Angles curves are
more L-shaped and indicate that more of a hot- spot problem
exists.
Figure VI.4 is a three dimensional plot of hazard versus
receptor location for the South Coast Air Basin. There are a
number of hot spots within the Air Basin. While each of the
peaks shown in Figure VI.4 represents a single source rather than
a multiple source problem, the peaks are caused by different
source types and different pollutants. Peak values Nos. 1 and 2
(in Orange County) are produced by "ethylene oxide emitters where
ethylene oxide is emitted by either food sterilizers or
laboratories. Peak value No.3 is caused primarily by a
ma"nufacturing facility which uses perchloroethylene in an
open-top vapor degreaser. Finally, peak value No.4 is caused by
a facility that emits benzene fugitives. Thus, air toxic hot
spots in the South Coast Air Basin are caused by a diverse mix of
source types and pollutants.
98
-------
HSUM
13. 11
\.0
\.0
8.74
4. 38
Figure VI. 2
HAZARDS FROM MAJOR POINT SOURCES
REGION=PHOENIX
3720
y
0.01
374.00
x
3690
440.00
-------
30
,-....
c.o
( 25
o
......
*"
"--/
...... (l) 20
o >
o (l)
-0
L
0 15
N
0
I
10
Figure VI.3
Key Air Taxies
Point
Source
Hazards
40
Los Angeles
35
5
o
10
20
30
40
50
60
70
80
90
100
[2:2J Base
Population (percentile)
[s:sJ NAAQS/NESHAPS
-------
Figure VI. 4
HAZARDS FROM MAJOR POINT SOURCES
REGION=LOS ANGELES
.HSUM
.... 303.31
o
....
3
2
1
1. 03
364.00
4
202.55
3790
101. 79
y
3746
384.67
x
3724
426.00
-------
3. Baton Rouge
Baton Rouge hot spot analysis results are presented in
Figure VI.5. Hazard levels from exposure to point source emitted
toxics are between those of Los Angeles and Phoenix but they are
expected to be reduced significantly under current regulatory
programs. For instance, Figure VI.5 shows that 10 percent of the
Baton Rouge population is exposed to a hazard of 1.1 x 10-5
or more in 1980 in Baton Rouge. In 1995, the highest6hazard
level in Baton Rouge drops to approximately 4.5 x 10- .
Figure VI.6 is a three dimensional plot of hazard versus
receptor location for East and West Baton Rouge Parishes. It
appears that downtown Baton Rouge sits in the saddle between two
areas of high hazard. As might be expected, the major chemical
manufacturing facilities in East Baton Rouge Parish are the
primary contributors to the two hot spots. A number of
pollutants contribute to the points of high hazard in Baton
Rouge. They include the following: carbon tetrachloride,
ethylene dichloride, benzene, perchloroethylene, trichloro-
ethylene, and vinyl chloride.
4. Baltimore
Baltimore has the highest hazard level from major point
sources of all the study areas. The point of maximum hazard in
Baltimore is at least an order of magnitude higher than that in
any of the other four cities. This high hazard level is caused
by chromium emissions from a refractory manufacturing plant and a
sodium chromate and sodium dichromate manufacturer.. Figure VI.?
shows that the hazard level is relatively constant for 90 percent
of the Baltimore population, but that a relatively small
percentge of the population is subjected to substantially higher
hazard. Unfortunately, current regulatory .programs are not
expected to reduce hazards associated with point source emitted
air toxics in Baltimore. .
Figure VI.8 shows the location of areas of maximum hazard
in Baltimore County. The primary peak is due to the chromium
emitted by the refractory manufacturing plant. The secondary
peak is a combination of chromium emissions from the refractory
and the sodium chromate and sodium dichromate manufacturer.
5. Philadelphia
Figure VI.9 shows the relationship between hazard level
and population for Philadelphia. Current regulations are
.expected to reduce the maximum hazard level in Philadelphia only
sl igh tly .
Figure VI.10 is a three dimensional plot of hazard versus
102
-------
Figure VI.5
Key Air Taxies Point Source Hazards
Baton Rouge
12
1 1
10
9
,--.,.
lD 8
(
0
...- 7
*"
'--'"
I-' Q) 6
0 >
w Q)
"'0 5
L
0
N 4
0
I
3
2
1
0
10 20 30 40 50 60 70 80 90 100
fZZ] Base
Population (percentile)
cs=sJ NAAQS/NESHAPS
-------
-Figure VI. 6
HAZARDS FROM MAJOR POINT SOURCES
HSUM
19.29
I-'
o
.l:'
13.08
6. 88
O. 67
. 656.00
REGION=BATON ROUGE
WEST. BATON ROUGE PARISH
EAST BATON ROUGE PARISH
3390.00
3376. 67
y
3363.33
x
3350.00
684.00
-------
~
o
\.J1
r--.
to
(
a
..- r--.
* en
,,--,IJ
c
-0
Wen
>:J
Wo
IJ-c
,,-I-
0"--'
N
o
I
2.1
2
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
o
Figure VI.7
,Key Air Taxies
Point
Source
Hazards
Baltimore
10
20
30
40
50
60
70
80
90
100
rz:zJ Base
Population (percentile)
cs:sJ NAAQS/NESHAPS
-------
HSUM
I-' 8381. 46.
o
0\
5599. 70
2817.95
Figure VI.8
l-IAZARDS FROM MAJOR POINT SOURCES
REGION=8ALTIMORE
4360.00
4353.33
36. 19 .
354.00
x
4340.00
385.00
-------
Figure VI.9
Key Air Taxies Point Source Hazards
Philadelphia
50
45
40
~ 35
to
(
0
~ 30
*
'--'"
....... Q) 25
0 >
~ Q)
-0 20
I...
0
N
0
I 15
10
5
0
10 20 30 40 50 60 70 80 90 100
IZZI Ba se
Popu1ation (percentile)
cs:sJ NAAQS/NESHAPS
-------
Figure VI. 10
HA.ZARDS FR011 MAJ-OR POINT SOURCES
REGION=PHILADELPHIA
HSUM
0.00
478.00
.
.
I-'
o
ex>
222.30
148.20
4440.00
74. 10
4431. 67
y
4423.33
x
4415.00
. 502.00
-------
receptor location for Philadelphia County. A POTW is the major
contributor to the point of maximum hazard shown in Figure VI.10.
The POTW emits a number of different air toxic compounds
including: formaldehyde, chloroform, perchloroethylene,
trichloroethylene, benzene, and vinyl chloride.
The secondary peak shown in Figure VI.10 is caused
primarily by chromium and nickel emissions from a fabricated
metals facility. It is also contributed to by the POTW.
C. SUMMARY
In summary, maximum hazards from major point sources vary
widely among the five study areas. The point source chromium
emissions in Baltimore are the major source of city-to-city
differences.
In most instances, it was found that risk hot spots are
caused by single sources, although these sources may emit more
than one air toxic compound. Thus, an analysis such as the one
reported here may be an effective tool for use by local agencies
in targeting control strategies toward those sources that produce
peak hazards.
109
-------
B(a)P
BACT
CAG
CDM
EDC
EPA
FMV CP
HEM
IFD
LAER
NAPAP
N ECR MP
NEDS
N ESHAPs
NSPS
NSR
Pechan
PIC
POTW.
PSD
PVC
RA-CT
RIM
ABBREVIATIONS AND ACRONYMS
Benzo(a)pyrene
Best Available Control Technology
Carcinogen Assessment Group
Climatological Dispersion Model
Ethylene dichloride
u.S. Environmental Protection Agency
Federal Motor Vehicle Control Program
Human Exposure Model
Industrial Facilities Discharge
Lowest Achievable Emission Rate
National Acid Precipitation Assessment Program
Northeast Corridor Regional Modeling Project
National Emissions Data System
National Emission Standards for Hazardous Air
Poll uta nt s
New Source Performance Standards
New Source Review
E.H. Pechan & Associates, Inc.
Products of incomplete combustion
Publicly Owned Treatment Works
Prevention of Significant Deterioration
Polyvinyl chloride
Reasonably Available Control Technology
Regulatory Impact Model
110
-------
SCAB
SCC
SIC
SIP
SHEAR
SMSA
STAR
TCE
TSP
UTM
VOC
ABBREVIATIONS AND ACRONYMS (continued)
South Coast Air Basin
Source Classification Code
Standard Industrial Classification
State Implementation Plans
Systems Applications Human Exposure and Risk (Model)
Standard Metropolitan Statistical Area
Stability Array
Trichloroethylene
Total suspended particulates
Universal Transverse Mercator
Volatile organic compounds
111
-------
REFERENCES
Automotive Engineering, 1984: "A Primer on Heavy Duty Diesel
Particulate Control," A.Y.t.Q1!1.Q.t..t..YJL.E.D,gj.D.,g.,g.rjlU~,' Vol. 92, No.
11, November 1984, pp. 63-70.
Benkov i tz, 1984: Ca rmen Benkov i tz, "U ncer tainty Analy si s of th e
NAPAP Inventory," (Draft), Brookhaven National Laboratory,
Upton, NY, May 1984.
Carey, 1981: Carey, Penny M., "Mobile Source Emissions of
Formaldehyde and Other Aldehydes," U.S. Environmental
Protection Agency, Ann Arbor, MI, May 1981.
DER, 1982: Department of Environmental Resources, "Pennsylvania
Residential Fuelwood Use Assessment, 1980-81," Office of
Resources Management, Bureau of Forestry, State of
Pennsylvania, Harrisburg, Pennsylvania, 1982.
EIA, 1982a:. Energy Information Administration, "Estimates of
U. S. Wood Energy Consumption from 1949 to 1981,"
DOE/EIA-0341, U.S. Department of Energy, Washington, DC,
1982.
EIA, 1982b:Energy Information Administration, "Residential
Energy Consumption Survey: Consumption and Expenditures,
April 1980-March 1981," DOE/EIA-0321/1, U. S. Department of
Energy, Washington, DC, 1982.
EIA, 1982c: Energy Information Administration, "Residential
Energy Consumption Survey: Housing Characteristics 1980,"
DOE/EIA-0314, U.S. Department of Energy, Washington, DC,
1 982 .
GCA, 1984: GCA Corporation, "Assessment of Chloroform Source
Categories, Final Report," GCA-TR-CH-84-12, prepared for U.S.
Environmental Protection Agency, Research Triangle Park, NC,
J ul y 1 984 .
Gray, 1982: "Mobil e Source Benz ene," U. S. Env ironmental
Protection Agency, Memorandum from Charles L. Gray, Director
of the Emission Control Technology Division, U.S.
Environmental Protection Agency, Ann Arbor, MI, to Donald R.
Goodw in, Di rector of the Emissi'on Standards, and Engi neering
Division, U.S. Environmental Protection Agency, Research
Triangle Park, NC, November 1982.
HID, 1984: Housing Industry Dynamics, "Sales and Marketing
Information for Members of the Wood Heating Alliance,"
Crofton, MD, 1984.
112
-------
IRS, 1983: Internal Revenue Service, "IRS Depreciation," U.S.
Department of the Treasury Publication No. 534, November
1983.
Jones, 1984:
1 984 .
L. Jones, "Benzene Fact Sheet," updated October
Lindgr en, 1978: L. H. Lindgr en, "Co st. Estimates for Emissi on
Control Related Components/Systems and Cost Methodology
Description," Rath and Strong, Inc., Lexington, MA, prepared
for U.S. Environmental Protection Agency, Ann Arbor, MI,
EPA-460/3-78-002, March 1978.
Marshall, 1981: Norman L. Marshall, "The Dynamics of Residential
Wood Energy Use in New England, 1970-2000," Policy Resource
Center, Thayer School of Engineering, Dartmouth College.
Hanover, New Hampshire, October 1981.
Metcalf and Eddy, 1979: Metcalf and Eddy, Inc., ~E~t~~Et~I
EDgjD~~I1Dg~__Ir~~tID~Dt~_Dj~~Q~El_ED~_~~~~, second
edition, revised by George Tchobanoglous, New York:
McGraw-Hill, 1979.
Metcalf and Eddy, 1981: Metcalf and Eddy, Inc., ~E~t~~~teI
EDgjD~~IjD~__~Qll~.Qli.QD_E.D.~L.f~ID~jDL.Q.r_Ji~~t~~Et~I , by
George Tchobanoglous, New. York: - McGraw-Hill, 1981.
Mohin, 1984: Tim Mohin, U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, "The Control of
Toxic Air Pollutants by Various Regulatory Strategies -
Draft," Research Triangle Park, NC, December 1984.
MRI, 1984: Midwest Research Institute, "Source Assessment of
Ethylene Dichloride Emissions, Final Report," MRI 7710-L,
Kansas City, MO, 1984.
NRBP, 1984: Northeast Regional Biomass Program, "Particulate
Emissions from Residential Wood Combustion - Summary Report,"
CONEG Policy Research Center, Inc., Washington, DC, May 1984.
NAS, 1983: National Academy of Sciences, Committee on pyrene and
Selected Analogues, National Research Council, "Polycyclic
Aromatic Hydrocarbons: Evaluation of Sources and Effects,"
National Academy Press, Washington, DC, 1983.
PEPCo, 1982: PEDCo Environmental, Inc., Krishnan, E.
G. Vinson Hellwig, "Trace Emissions from Coal and
Combustion" in £nxjI.QnID~n~~l_.fIQgI~~, VOl.1, no.
November 1982.
Radha and
Oil
4,
113
-------
Radian, 1984a: Radian Corporation, "Study of Sources of
Chromium, Nickel and Manganese Air Emissions, Final Report,"
prepared for Pollutant Assessment Branch, Office of Air
Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC, May 1984.
Radian, 1984b: Radian Corporation, "R~sk Estimates/Secondary
Lead Smel ting," ~emorandum from Radi,an Corporation to
W. Peters et al., U.S. Environmental Protection Agency,
Research Triangle Park, NC, October 1984.
SAI, 1982: Systems Applications, Inc., "Human Exposure to
Atmospheric Concentrations of Selected Chemicals," prepared
for Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, Research Triangle Park, NC,
February 1982.
SAI, 1983: Sy stems Appl ica ti ons, Inc., "U ser r S Ma nual for
SHEAR," San Rafael, CA, May 1983.
SCAQMD, 1983: South Coast Air Quality Management District,
"Emissions of Potentially Toxic/Hazardous Air Contaminants in
the South Coast Air Basin," El Monte, CA, September 1983.
Skog and Watterson, 1983: Kenneth E. Skog and Irene W~tterson,
"Residential Fuelwood in the United States: 1980-1981,ri .
Forest Products Laboratory, U.S. Forest Service, Madison, WI,
Jul y 1 983 . . ..
USDA, 1980: U.S. Department of Agriculture, "Prospectus:
Firewood Manufacturing and Marketing," U. S. Forest Service
NA-FR-17, Madison, WI, February 1980.
U.S. DOC, 1975: U.S. Department of Commerce, Bureau of the
Census, "Construction Reports: Characteristics of New
Housing," C 25-74-13, Washington, DC, August 1975.
U.S. DOC, 1980: U.S. Department of Commerce, Bureau of the
Census, "Construction Reports: Characteristics of New
Housing," C 25-79-13, Washington, DC, August 1980.
U.S. DOC, 1982a: U.S. Department of Commerce, Bureau of the
Census. "1980 Decennial Census, Detailed Housing
Characteristics," Washington, DC, 1982.
U.S. DOC, 1982b: U.S. Department of Commerce, Bureau of the
. Census, "Construction Reports: Characteristics of New
. Housing," C 25-81-13, Washington, DC, August 1982.
114
-------
u.s. DOC, 1983a: U.S. Department of Commerce, Bureau of the
Census, "1980 Annual Housing Survey: Energy Related Housing
C h a r act e r i s tic sPa r t F," H - 1 50- 80, Wash in g ton, DC: 1 983 .
u.S. DOC, 1983b: U.S. Department of Commerce, Bureau of the
Census, ~Q~nt~_~n~_~j~~_Ds~~_~QQ~_1~a1, Washington, DC:
U.S. Government Printing Office, 1983.
u.S. DOL, 1958: u.S. Department of Labor, Bureau of Labor
Sta ti sti cs, "N ew Housi ng and its Ma teri al s: 1940-1956, "
Bulletin No. 1231, Washington, DC, 1958.
u.S. EPA, 1979a: U.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, "1976 National Emissions
Report," EPA-450/4-79-019, Research Triangle Park, NC, August
1 97 9 .
u.S. EPA, 1979b: u.S. Environmental Protection Agency, Office of
Mobile Source Air Pollution Control, "Regulatory Analysis and
Environmental Impact of Final Emission Regulations for 1984
and Later Model Heavy Duty Engi ne s," Ann Arbor, MI, -December
1 1, 1 97 9 .
u.S. EPA, 1980a: U.S. Environmental Protection Agency, "Organic
Chemical Manufacturing Volume 5: Absorption, Condensation,
and Absorption Techniques," EPA-450/3-80-027, December 1980.
U. S. EPA, 1 980b : u. S. Env ironm ental Pr otecti on Age ncy, "0 rga ni c
.Chemical Manufacturing Volume 4: Combustion Control
Devices," EPA-450/3-80-026, December 1980.
u. S. EPA, 1 982 a: U. S. Env ironm ental Pr otecti on Agency, "V iny 1
Chloride - A Rev iew of National Emissions Standards~" EPA-
450/3-82-003, Research Triangle Park, NC, February 1982.
u.S. EPA, 1982b~ U.S. Environmental Protection Agency, "The Cost
Digest," May 1982.
u.S. EPA, 1982c: U.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, "Benzene Fugitive
Emissions - Background Information for Promulgated
Standards," EPA-450/3-80-032b, Research Triangle Park, NC,
June 1982.
u.S. EPA, 1982d: U.s. Environmental Protection Agency, Office of
Air Quality Planning and Standards, "Preliminary Study of.
. Sources of Inorganic Arsenic," EPA-450/5-82-005, Research
Triangle Park, NC, August 1982.
115
-------
u.s. EPA, 1982e: u.s. Environmental Protection Agency, "Cont'rol
Techniques for Particulate Emissions from Stationary
Sources," Vol. 1, EPA-450/3-81-005a, September 1982.
u.S. EPA, 1983: u.S. Environmental Protection Agency, "Inorganic
Arsenic from Glass Manufacturing Plants - Background
Information for Proposed Standards - Draft EIS," EPA-
450/3-83-011a, Research Triangle Park, NC, March 1983..
u.S. EPA, 1984a: u.S. Environmental Protection Agency,
"Compilation of Air Pollutant Emission Factors, Third Edition
(including Supplements 1-15) ," AP-42, January 1984.
u.S. EPA, 1984b: u.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, "Locating and Estimating
Air Emissions from Sources of Carbon Tetrachloride," .
EPA-450/4-84-007b, Research Triangle Park, NC, March 1984.
u.S. EPA, 1984c: u.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, "Locating and Estimating
Air Emissions .[.rom Sources of Chloroform," EPA-450/4-84-007c,
Research Triangle Park, NC, March 1984.
u.S. EPA, 1984d: u.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, "Locating and Estimating
Air Emissions from Sources of Ethylene Dichloride,"
EPA-450/4-84-007d, Research Triangle Park, NC, March 1984.
u. S. EPA, 1 984e: . u. S. Env ironmental Pr otecti on Age ncy, Office of
. Air Quality Planning and Standards, "Locating and Estimating
Air Emissions from Sources of Formaldehyde," EPA-
450/4-84-007e, . Research Triangle Park, NC, March 1984.
U. S. EPA, 1984 f: U. S. Env ironmental Pr otecti on Age ncy, "Loca ti ng
and Estimating Air Emissions from Sources of Nickel," EPA-
450/4-84-007f, Research Triangle Park, NC, March 1984.
u.S. EPA, 1984g: u.S. Environmental Protection Agency, "Size
Specific Total Particulate Emission Factors for Mobile
Sources," EPA-TSS-PA-84-1, Ann Arbor, MI, April 1984.
u.S. EPA, 1984h: u.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, "Benzene Emissions from
Coke By-product Recovery Plants - Background Information for
Proposed Standards," EPA-450/3-83-016a, Draft EIS, Research
Triangle Park, NC, Mqy 1984.
u.s. EPA, 1984i: u.s. Environmental Protection Agency,
"Preliminary Study of Sources of Ethylene Oxide, Draft
Report," Research Triangle Park, NC, June 1984.
116
-------
u.s. EPA, 1984j: u.s. Environmental Protection Agency, Office of
Air Quality Planning and Standar.ds, "Benzene Emissions From
Nitrobenzene, Chlorobenzene, Linear Alkylbenzene and Ethylene
Plant Vents - Background Information~" EPA-450/5-84-002,
Research Triangle Park, ~C, July 1984.
u.s. EPA, 1984k: u.s. Environmental Protection Agency, Office of
Air Quality Planning and Standards and Office of Mobile
Sources, "Evaluation of Air Pollution Regulatory Strategies
for Gasoline Marketing Industry," EPA-450/3-84-012a & b,
Research Triangle Park, NC, July 1984.
u.S. EPA, 19841: u.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, "Locating and Estimating
Air Emissions from Sources of Chromium," EPA-450/4-84-007g,
Research Triangle Park, NC, July 1984.
U. S. EPA, 1984m: U. S. Env ironmental Protection Agency, "Draft
Regulatory Impact Analysis of Proposal Rules Limiting the
Lead Content of Gasoline," Washington, DC, July 23, 1984.
U. S. EPA, 1985: U. S. Env ironmental Protecti on Agency, "The
Control of Toxic Air Pollutants by Various Regulatory
Strategies," T. J. Mohin and R. M. Schell, March 1985.
Versar, 1984: Versar Inc., "Hazardous Air Pollutants, A
Prel iminary Exposure and Risk Appraisal for 35 U. S.
Counties," Springfield, VA, September 1984.
117
------- |